1046 lines
28 KiB
JSON
1046 lines
28 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 0.41841004184100417,
|
|
"eval_steps": 500,
|
|
"global_step": 500,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"entropy": 10.742597675323486,
|
|
"epoch": 0.0041841004184100415,
|
|
"grad_norm": 5.46875,
|
|
"learning_rate": 2e-06,
|
|
"loss": 10.7608,
|
|
"mean_token_accuracy": 9.661835501901805e-05,
|
|
"num_tokens": 9848.0,
|
|
"step": 5
|
|
},
|
|
{
|
|
"entropy": 10.74262809753418,
|
|
"epoch": 0.008368200836820083,
|
|
"grad_norm": 5.5,
|
|
"learning_rate": 4.5e-06,
|
|
"loss": 10.723,
|
|
"mean_token_accuracy": 0.0,
|
|
"num_tokens": 18196.0,
|
|
"step": 10
|
|
},
|
|
{
|
|
"entropy": 10.742603778839111,
|
|
"epoch": 0.012552301255230125,
|
|
"grad_norm": 5.34375,
|
|
"learning_rate": 7e-06,
|
|
"loss": 10.703,
|
|
"mean_token_accuracy": 0.0,
|
|
"num_tokens": 26927.0,
|
|
"step": 15
|
|
},
|
|
{
|
|
"entropy": 10.742668437957764,
|
|
"epoch": 0.016736401673640166,
|
|
"grad_norm": 5.65625,
|
|
"learning_rate": 9.5e-06,
|
|
"loss": 10.6357,
|
|
"mean_token_accuracy": 0.0005675206426531077,
|
|
"num_tokens": 34987.0,
|
|
"step": 20
|
|
},
|
|
{
|
|
"entropy": 10.742566776275634,
|
|
"epoch": 0.02092050209205021,
|
|
"grad_norm": 5.03125,
|
|
"learning_rate": 1.2e-05,
|
|
"loss": 10.5588,
|
|
"mean_token_accuracy": 0.010127833811566234,
|
|
"num_tokens": 44936.0,
|
|
"step": 25
|
|
},
|
|
{
|
|
"entropy": 10.742234230041504,
|
|
"epoch": 0.02510460251046025,
|
|
"grad_norm": 4.3125,
|
|
"learning_rate": 1.4500000000000002e-05,
|
|
"loss": 10.427,
|
|
"mean_token_accuracy": 0.0348518967628479,
|
|
"num_tokens": 54562.0,
|
|
"step": 30
|
|
},
|
|
{
|
|
"entropy": 10.741086196899413,
|
|
"epoch": 0.029288702928870293,
|
|
"grad_norm": 3.359375,
|
|
"learning_rate": 1.7000000000000003e-05,
|
|
"loss": 10.2814,
|
|
"mean_token_accuracy": 0.03771395143121481,
|
|
"num_tokens": 63844.0,
|
|
"step": 35
|
|
},
|
|
{
|
|
"entropy": 10.738179874420165,
|
|
"epoch": 0.03347280334728033,
|
|
"grad_norm": 2.734375,
|
|
"learning_rate": 1.95e-05,
|
|
"loss": 10.1587,
|
|
"mean_token_accuracy": 0.04109239801764488,
|
|
"num_tokens": 74331.0,
|
|
"step": 40
|
|
},
|
|
{
|
|
"entropy": 10.734289169311523,
|
|
"epoch": 0.03765690376569038,
|
|
"grad_norm": 2.484375,
|
|
"learning_rate": 2.2e-05,
|
|
"loss": 9.995,
|
|
"mean_token_accuracy": 0.04218972772359848,
|
|
"num_tokens": 83028.0,
|
|
"step": 45
|
|
},
|
|
{
|
|
"entropy": 10.730850219726562,
|
|
"epoch": 0.04184100418410042,
|
|
"grad_norm": 2.15625,
|
|
"learning_rate": 2.4500000000000003e-05,
|
|
"loss": 9.9353,
|
|
"mean_token_accuracy": 0.040206428803503516,
|
|
"num_tokens": 91395.0,
|
|
"step": 50
|
|
},
|
|
{
|
|
"entropy": 10.729630088806152,
|
|
"epoch": 0.04602510460251046,
|
|
"grad_norm": 2.28125,
|
|
"learning_rate": 2.7e-05,
|
|
"loss": 9.8674,
|
|
"mean_token_accuracy": 0.048422401398420335,
|
|
"num_tokens": 101868.0,
|
|
"step": 55
|
|
},
|
|
{
|
|
"entropy": 10.727583599090575,
|
|
"epoch": 0.0502092050209205,
|
|
"grad_norm": 2.015625,
|
|
"learning_rate": 2.95e-05,
|
|
"loss": 9.7933,
|
|
"mean_token_accuracy": 0.04613714665174484,
|
|
"num_tokens": 110403.0,
|
|
"step": 60
|
|
},
|
|
{
|
|
"entropy": 10.723635578155518,
|
|
"epoch": 0.05439330543933055,
|
|
"grad_norm": 1.921875,
|
|
"learning_rate": 3.2e-05,
|
|
"loss": 9.8021,
|
|
"mean_token_accuracy": 0.041691455990076065,
|
|
"num_tokens": 119557.0,
|
|
"step": 65
|
|
},
|
|
{
|
|
"entropy": 10.719999694824219,
|
|
"epoch": 0.058577405857740586,
|
|
"grad_norm": 2.0,
|
|
"learning_rate": 3.4500000000000005e-05,
|
|
"loss": 9.676,
|
|
"mean_token_accuracy": 0.04646212197840214,
|
|
"num_tokens": 128903.0,
|
|
"step": 70
|
|
},
|
|
{
|
|
"entropy": 10.71647481918335,
|
|
"epoch": 0.06276150627615062,
|
|
"grad_norm": 1.8359375,
|
|
"learning_rate": 3.7e-05,
|
|
"loss": 9.6628,
|
|
"mean_token_accuracy": 0.041718531772494315,
|
|
"num_tokens": 137630.0,
|
|
"step": 75
|
|
},
|
|
{
|
|
"entropy": 10.713119316101075,
|
|
"epoch": 0.06694560669456066,
|
|
"grad_norm": 2.0,
|
|
"learning_rate": 3.95e-05,
|
|
"loss": 9.5313,
|
|
"mean_token_accuracy": 0.0421627514064312,
|
|
"num_tokens": 146335.0,
|
|
"step": 80
|
|
},
|
|
{
|
|
"entropy": 10.708419799804688,
|
|
"epoch": 0.07112970711297072,
|
|
"grad_norm": 2.0,
|
|
"learning_rate": 4.2000000000000004e-05,
|
|
"loss": 9.4686,
|
|
"mean_token_accuracy": 0.04617839008569717,
|
|
"num_tokens": 155766.0,
|
|
"step": 85
|
|
},
|
|
{
|
|
"entropy": 10.702109909057617,
|
|
"epoch": 0.07531380753138076,
|
|
"grad_norm": 1.859375,
|
|
"learning_rate": 4.45e-05,
|
|
"loss": 9.412,
|
|
"mean_token_accuracy": 0.04476847723126411,
|
|
"num_tokens": 165844.0,
|
|
"step": 90
|
|
},
|
|
{
|
|
"entropy": 10.693535137176514,
|
|
"epoch": 0.0794979079497908,
|
|
"grad_norm": 1.90625,
|
|
"learning_rate": 4.7000000000000004e-05,
|
|
"loss": 9.3394,
|
|
"mean_token_accuracy": 0.05125464424490929,
|
|
"num_tokens": 174858.0,
|
|
"step": 95
|
|
},
|
|
{
|
|
"entropy": 10.678814315795899,
|
|
"epoch": 0.08368200836820083,
|
|
"grad_norm": 1.9140625,
|
|
"learning_rate": 4.9500000000000004e-05,
|
|
"loss": 9.2166,
|
|
"mean_token_accuracy": 0.06135832369327545,
|
|
"num_tokens": 184020.0,
|
|
"step": 100
|
|
},
|
|
{
|
|
"entropy": 10.653692531585694,
|
|
"epoch": 0.08786610878661087,
|
|
"grad_norm": 1.8828125,
|
|
"learning_rate": 5.2e-05,
|
|
"loss": 9.115,
|
|
"mean_token_accuracy": 0.05314192697405815,
|
|
"num_tokens": 193714.0,
|
|
"step": 105
|
|
},
|
|
{
|
|
"entropy": 10.62697982788086,
|
|
"epoch": 0.09205020920502092,
|
|
"grad_norm": 1.90625,
|
|
"learning_rate": 5.45e-05,
|
|
"loss": 9.0408,
|
|
"mean_token_accuracy": 0.05419124960899353,
|
|
"num_tokens": 203139.0,
|
|
"step": 110
|
|
},
|
|
{
|
|
"entropy": 10.581832027435302,
|
|
"epoch": 0.09623430962343096,
|
|
"grad_norm": 1.875,
|
|
"learning_rate": 5.7e-05,
|
|
"loss": 8.8865,
|
|
"mean_token_accuracy": 0.05798161998391151,
|
|
"num_tokens": 212446.0,
|
|
"step": 115
|
|
},
|
|
{
|
|
"entropy": 10.510433387756347,
|
|
"epoch": 0.100418410041841,
|
|
"grad_norm": 1.8046875,
|
|
"learning_rate": 5.9499999999999996e-05,
|
|
"loss": 8.7392,
|
|
"mean_token_accuracy": 0.06025223545730114,
|
|
"num_tokens": 220910.0,
|
|
"step": 120
|
|
},
|
|
{
|
|
"entropy": 10.451266765594482,
|
|
"epoch": 0.10460251046025104,
|
|
"grad_norm": 1.6953125,
|
|
"learning_rate": 6.2e-05,
|
|
"loss": 8.6729,
|
|
"mean_token_accuracy": 0.055778010934591296,
|
|
"num_tokens": 230522.0,
|
|
"step": 125
|
|
},
|
|
{
|
|
"entropy": 10.406130599975587,
|
|
"epoch": 0.1087866108786611,
|
|
"grad_norm": 1.7578125,
|
|
"learning_rate": 6.450000000000001e-05,
|
|
"loss": 8.5065,
|
|
"mean_token_accuracy": 0.05544214770197868,
|
|
"num_tokens": 240092.0,
|
|
"step": 130
|
|
},
|
|
{
|
|
"entropy": 10.331532001495361,
|
|
"epoch": 0.11297071129707113,
|
|
"grad_norm": 1.5625,
|
|
"learning_rate": 6.7e-05,
|
|
"loss": 8.4083,
|
|
"mean_token_accuracy": 0.06165078543126583,
|
|
"num_tokens": 249478.0,
|
|
"step": 135
|
|
},
|
|
{
|
|
"entropy": 10.237152481079102,
|
|
"epoch": 0.11715481171548117,
|
|
"grad_norm": 1.71875,
|
|
"learning_rate": 6.950000000000001e-05,
|
|
"loss": 8.2,
|
|
"mean_token_accuracy": 0.06240398585796356,
|
|
"num_tokens": 257714.0,
|
|
"step": 140
|
|
},
|
|
{
|
|
"entropy": 10.116984844207764,
|
|
"epoch": 0.12133891213389121,
|
|
"grad_norm": 1.4296875,
|
|
"learning_rate": 7.2e-05,
|
|
"loss": 8.1928,
|
|
"mean_token_accuracy": 0.05669947192072868,
|
|
"num_tokens": 266569.0,
|
|
"step": 145
|
|
},
|
|
{
|
|
"entropy": 10.02002773284912,
|
|
"epoch": 0.12552301255230125,
|
|
"grad_norm": 1.375,
|
|
"learning_rate": 7.45e-05,
|
|
"loss": 8.0619,
|
|
"mean_token_accuracy": 0.06153124496340752,
|
|
"num_tokens": 276688.0,
|
|
"step": 150
|
|
},
|
|
{
|
|
"entropy": 9.869840335845947,
|
|
"epoch": 0.1297071129707113,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 7.7e-05,
|
|
"loss": 7.8676,
|
|
"mean_token_accuracy": 0.06409889869391919,
|
|
"num_tokens": 285430.0,
|
|
"step": 155
|
|
},
|
|
{
|
|
"entropy": 9.67565851211548,
|
|
"epoch": 0.13389121338912133,
|
|
"grad_norm": 1.4375,
|
|
"learning_rate": 7.950000000000001e-05,
|
|
"loss": 7.789,
|
|
"mean_token_accuracy": 0.06454115808010101,
|
|
"num_tokens": 295727.0,
|
|
"step": 160
|
|
},
|
|
{
|
|
"entropy": 9.464604949951172,
|
|
"epoch": 0.13807531380753138,
|
|
"grad_norm": 1.1171875,
|
|
"learning_rate": 8.2e-05,
|
|
"loss": 7.6957,
|
|
"mean_token_accuracy": 0.06615287810564041,
|
|
"num_tokens": 304854.0,
|
|
"step": 165
|
|
},
|
|
{
|
|
"entropy": 9.24670057296753,
|
|
"epoch": 0.14225941422594143,
|
|
"grad_norm": 1.3046875,
|
|
"learning_rate": 8.450000000000001e-05,
|
|
"loss": 7.6093,
|
|
"mean_token_accuracy": 0.07150843031704426,
|
|
"num_tokens": 313579.0,
|
|
"step": 170
|
|
},
|
|
{
|
|
"entropy": 9.048341655731202,
|
|
"epoch": 0.14644351464435146,
|
|
"grad_norm": 1.0234375,
|
|
"learning_rate": 8.7e-05,
|
|
"loss": 7.6128,
|
|
"mean_token_accuracy": 0.07077031470835209,
|
|
"num_tokens": 323310.0,
|
|
"step": 175
|
|
},
|
|
{
|
|
"entropy": 8.823628616333007,
|
|
"epoch": 0.1506276150627615,
|
|
"grad_norm": 1.03125,
|
|
"learning_rate": 8.95e-05,
|
|
"loss": 7.4519,
|
|
"mean_token_accuracy": 0.06803311333060265,
|
|
"num_tokens": 331720.0,
|
|
"step": 180
|
|
},
|
|
{
|
|
"entropy": 8.658255577087402,
|
|
"epoch": 0.15481171548117154,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 9.2e-05,
|
|
"loss": 7.3797,
|
|
"mean_token_accuracy": 0.07455825209617614,
|
|
"num_tokens": 340331.0,
|
|
"step": 185
|
|
},
|
|
{
|
|
"entropy": 8.487533950805664,
|
|
"epoch": 0.1589958158995816,
|
|
"grad_norm": 1.2421875,
|
|
"learning_rate": 9.45e-05,
|
|
"loss": 7.3713,
|
|
"mean_token_accuracy": 0.07890852093696595,
|
|
"num_tokens": 349677.0,
|
|
"step": 190
|
|
},
|
|
{
|
|
"entropy": 8.328230381011963,
|
|
"epoch": 0.16317991631799164,
|
|
"grad_norm": 1.3046875,
|
|
"learning_rate": 9.7e-05,
|
|
"loss": 7.3358,
|
|
"mean_token_accuracy": 0.06953225210309029,
|
|
"num_tokens": 358674.0,
|
|
"step": 195
|
|
},
|
|
{
|
|
"entropy": 8.279506969451905,
|
|
"epoch": 0.16736401673640167,
|
|
"grad_norm": 0.9609375,
|
|
"learning_rate": 9.95e-05,
|
|
"loss": 7.33,
|
|
"mean_token_accuracy": 0.07827548906207085,
|
|
"num_tokens": 368258.0,
|
|
"step": 200
|
|
},
|
|
{
|
|
"entropy": 8.202318572998047,
|
|
"epoch": 0.17154811715481172,
|
|
"grad_norm": 0.98828125,
|
|
"learning_rate": 0.000102,
|
|
"loss": 7.2291,
|
|
"mean_token_accuracy": 0.08247470743954181,
|
|
"num_tokens": 377083.0,
|
|
"step": 205
|
|
},
|
|
{
|
|
"entropy": 8.081873416900635,
|
|
"epoch": 0.17573221757322174,
|
|
"grad_norm": 1.0859375,
|
|
"learning_rate": 0.00010449999999999999,
|
|
"loss": 7.2799,
|
|
"mean_token_accuracy": 0.07478422187268734,
|
|
"num_tokens": 385456.0,
|
|
"step": 210
|
|
},
|
|
{
|
|
"entropy": 8.072468662261963,
|
|
"epoch": 0.1799163179916318,
|
|
"grad_norm": 1.46875,
|
|
"learning_rate": 0.000107,
|
|
"loss": 7.1697,
|
|
"mean_token_accuracy": 0.0813635230064392,
|
|
"num_tokens": 395295.0,
|
|
"step": 215
|
|
},
|
|
{
|
|
"entropy": 8.06587061882019,
|
|
"epoch": 0.18410041841004185,
|
|
"grad_norm": 1.2109375,
|
|
"learning_rate": 0.0001095,
|
|
"loss": 7.2016,
|
|
"mean_token_accuracy": 0.07997569367289543,
|
|
"num_tokens": 405023.0,
|
|
"step": 220
|
|
},
|
|
{
|
|
"entropy": 7.917844343185425,
|
|
"epoch": 0.18828451882845187,
|
|
"grad_norm": 0.9140625,
|
|
"learning_rate": 0.000112,
|
|
"loss": 7.2075,
|
|
"mean_token_accuracy": 0.07806592285633088,
|
|
"num_tokens": 414224.0,
|
|
"step": 225
|
|
},
|
|
{
|
|
"entropy": 7.906620788574219,
|
|
"epoch": 0.19246861924686193,
|
|
"grad_norm": 0.9375,
|
|
"learning_rate": 0.0001145,
|
|
"loss": 7.0979,
|
|
"mean_token_accuracy": 0.07944381982088089,
|
|
"num_tokens": 423951.0,
|
|
"step": 230
|
|
},
|
|
{
|
|
"entropy": 7.995381021499634,
|
|
"epoch": 0.19665271966527198,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 0.00011700000000000001,
|
|
"loss": 7.1259,
|
|
"mean_token_accuracy": 0.08254904597997666,
|
|
"num_tokens": 432427.0,
|
|
"step": 235
|
|
},
|
|
{
|
|
"entropy": 7.85433292388916,
|
|
"epoch": 0.200836820083682,
|
|
"grad_norm": 1.1171875,
|
|
"learning_rate": 0.00011949999999999999,
|
|
"loss": 7.0351,
|
|
"mean_token_accuracy": 0.08393027260899544,
|
|
"num_tokens": 441560.0,
|
|
"step": 240
|
|
},
|
|
{
|
|
"entropy": 7.8878974437713625,
|
|
"epoch": 0.20502092050209206,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 0.000122,
|
|
"loss": 7.1383,
|
|
"mean_token_accuracy": 0.0793025080114603,
|
|
"num_tokens": 450364.0,
|
|
"step": 245
|
|
},
|
|
{
|
|
"entropy": 7.774558973312378,
|
|
"epoch": 0.20920502092050208,
|
|
"grad_norm": 0.9140625,
|
|
"learning_rate": 0.0001245,
|
|
"loss": 7.0707,
|
|
"mean_token_accuracy": 0.08014305979013443,
|
|
"num_tokens": 460076.0,
|
|
"step": 250
|
|
},
|
|
{
|
|
"entropy": 7.792851877212525,
|
|
"epoch": 0.21338912133891214,
|
|
"grad_norm": 0.94921875,
|
|
"learning_rate": 0.000127,
|
|
"loss": 7.023,
|
|
"mean_token_accuracy": 0.08774993643164634,
|
|
"num_tokens": 468718.0,
|
|
"step": 255
|
|
},
|
|
{
|
|
"entropy": 7.835711526870727,
|
|
"epoch": 0.2175732217573222,
|
|
"grad_norm": 1.1484375,
|
|
"learning_rate": 0.0001295,
|
|
"loss": 7.0841,
|
|
"mean_token_accuracy": 0.08281839713454246,
|
|
"num_tokens": 478436.0,
|
|
"step": 260
|
|
},
|
|
{
|
|
"entropy": 7.807744693756104,
|
|
"epoch": 0.2217573221757322,
|
|
"grad_norm": 1.3046875,
|
|
"learning_rate": 0.000132,
|
|
"loss": 7.0932,
|
|
"mean_token_accuracy": 0.08015878945589065,
|
|
"num_tokens": 486833.0,
|
|
"step": 265
|
|
},
|
|
{
|
|
"entropy": 7.68460259437561,
|
|
"epoch": 0.22594142259414227,
|
|
"grad_norm": 1.0546875,
|
|
"learning_rate": 0.00013450000000000002,
|
|
"loss": 6.9589,
|
|
"mean_token_accuracy": 0.09470265135169029,
|
|
"num_tokens": 495119.0,
|
|
"step": 270
|
|
},
|
|
{
|
|
"entropy": 7.6987611770629885,
|
|
"epoch": 0.2301255230125523,
|
|
"grad_norm": 1.515625,
|
|
"learning_rate": 0.00013700000000000002,
|
|
"loss": 7.1042,
|
|
"mean_token_accuracy": 0.09113300144672394,
|
|
"num_tokens": 505396.0,
|
|
"step": 275
|
|
},
|
|
{
|
|
"entropy": 7.621922779083252,
|
|
"epoch": 0.23430962343096234,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.0001395,
|
|
"loss": 6.9717,
|
|
"mean_token_accuracy": 0.08859988525509835,
|
|
"num_tokens": 515259.0,
|
|
"step": 280
|
|
},
|
|
{
|
|
"entropy": 7.760473251342773,
|
|
"epoch": 0.2384937238493724,
|
|
"grad_norm": 0.99609375,
|
|
"learning_rate": 0.00014199999999999998,
|
|
"loss": 6.9769,
|
|
"mean_token_accuracy": 0.09006874561309815,
|
|
"num_tokens": 523608.0,
|
|
"step": 285
|
|
},
|
|
{
|
|
"entropy": 7.605395030975342,
|
|
"epoch": 0.24267782426778242,
|
|
"grad_norm": 1.109375,
|
|
"learning_rate": 0.0001445,
|
|
"loss": 6.9736,
|
|
"mean_token_accuracy": 0.08525220081210136,
|
|
"num_tokens": 534613.0,
|
|
"step": 290
|
|
},
|
|
{
|
|
"entropy": 7.54643816947937,
|
|
"epoch": 0.24686192468619247,
|
|
"grad_norm": 1.015625,
|
|
"learning_rate": 0.000147,
|
|
"loss": 6.9884,
|
|
"mean_token_accuracy": 0.09405484721064568,
|
|
"num_tokens": 543347.0,
|
|
"step": 295
|
|
},
|
|
{
|
|
"entropy": 7.628460121154785,
|
|
"epoch": 0.2510460251046025,
|
|
"grad_norm": 0.984375,
|
|
"learning_rate": 0.0001495,
|
|
"loss": 6.956,
|
|
"mean_token_accuracy": 0.08765949010848999,
|
|
"num_tokens": 552371.0,
|
|
"step": 300
|
|
},
|
|
{
|
|
"entropy": 7.5912059307098385,
|
|
"epoch": 0.25523012552301255,
|
|
"grad_norm": 1.0625,
|
|
"learning_rate": 0.000152,
|
|
"loss": 6.8822,
|
|
"mean_token_accuracy": 0.08999307677149773,
|
|
"num_tokens": 560742.0,
|
|
"step": 305
|
|
},
|
|
{
|
|
"entropy": 7.553208541870117,
|
|
"epoch": 0.2594142259414226,
|
|
"grad_norm": 1.1015625,
|
|
"learning_rate": 0.00015450000000000001,
|
|
"loss": 6.9822,
|
|
"mean_token_accuracy": 0.09437551945447922,
|
|
"num_tokens": 569960.0,
|
|
"step": 310
|
|
},
|
|
{
|
|
"entropy": 7.580251216888428,
|
|
"epoch": 0.26359832635983266,
|
|
"grad_norm": 0.9140625,
|
|
"learning_rate": 0.000157,
|
|
"loss": 6.9347,
|
|
"mean_token_accuracy": 0.08774123415350914,
|
|
"num_tokens": 578859.0,
|
|
"step": 315
|
|
},
|
|
{
|
|
"entropy": 7.558048486709595,
|
|
"epoch": 0.26778242677824265,
|
|
"grad_norm": 1.0078125,
|
|
"learning_rate": 0.0001595,
|
|
"loss": 6.9406,
|
|
"mean_token_accuracy": 0.09432831630110741,
|
|
"num_tokens": 588146.0,
|
|
"step": 320
|
|
},
|
|
{
|
|
"entropy": 7.445785140991211,
|
|
"epoch": 0.2719665271966527,
|
|
"grad_norm": 1.1015625,
|
|
"learning_rate": 0.000162,
|
|
"loss": 6.8822,
|
|
"mean_token_accuracy": 0.0896220214664936,
|
|
"num_tokens": 597268.0,
|
|
"step": 325
|
|
},
|
|
{
|
|
"entropy": 7.3907698631286625,
|
|
"epoch": 0.27615062761506276,
|
|
"grad_norm": 1.0,
|
|
"learning_rate": 0.00016450000000000001,
|
|
"loss": 6.8154,
|
|
"mean_token_accuracy": 0.0945595309138298,
|
|
"num_tokens": 606199.0,
|
|
"step": 330
|
|
},
|
|
{
|
|
"entropy": 7.577031898498535,
|
|
"epoch": 0.2803347280334728,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 0.00016700000000000002,
|
|
"loss": 6.9307,
|
|
"mean_token_accuracy": 0.08274017758667469,
|
|
"num_tokens": 615966.0,
|
|
"step": 335
|
|
},
|
|
{
|
|
"entropy": 7.524800825119018,
|
|
"epoch": 0.28451882845188287,
|
|
"grad_norm": 1.0625,
|
|
"learning_rate": 0.00016950000000000003,
|
|
"loss": 6.9039,
|
|
"mean_token_accuracy": 0.08914825171232224,
|
|
"num_tokens": 625689.0,
|
|
"step": 340
|
|
},
|
|
{
|
|
"entropy": 7.424820232391357,
|
|
"epoch": 0.28870292887029286,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.00017199999999999998,
|
|
"loss": 6.9367,
|
|
"mean_token_accuracy": 0.08863812088966369,
|
|
"num_tokens": 634877.0,
|
|
"step": 345
|
|
},
|
|
{
|
|
"entropy": 7.479735708236694,
|
|
"epoch": 0.2928870292887029,
|
|
"grad_norm": 1.125,
|
|
"learning_rate": 0.00017449999999999999,
|
|
"loss": 6.8576,
|
|
"mean_token_accuracy": 0.09228913635015487,
|
|
"num_tokens": 644042.0,
|
|
"step": 350
|
|
},
|
|
{
|
|
"entropy": 7.3995614528656,
|
|
"epoch": 0.29707112970711297,
|
|
"grad_norm": 1.1796875,
|
|
"learning_rate": 0.000177,
|
|
"loss": 6.8397,
|
|
"mean_token_accuracy": 0.09121191054582596,
|
|
"num_tokens": 653072.0,
|
|
"step": 355
|
|
},
|
|
{
|
|
"entropy": 7.390188837051392,
|
|
"epoch": 0.301255230125523,
|
|
"grad_norm": 0.9453125,
|
|
"learning_rate": 0.0001795,
|
|
"loss": 6.8429,
|
|
"mean_token_accuracy": 0.0936873622238636,
|
|
"num_tokens": 663686.0,
|
|
"step": 360
|
|
},
|
|
{
|
|
"entropy": 7.433347082138061,
|
|
"epoch": 0.3054393305439331,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.000182,
|
|
"loss": 6.7824,
|
|
"mean_token_accuracy": 0.09806219711899758,
|
|
"num_tokens": 672293.0,
|
|
"step": 365
|
|
},
|
|
{
|
|
"entropy": 7.224409723281861,
|
|
"epoch": 0.30962343096234307,
|
|
"grad_norm": 1.046875,
|
|
"learning_rate": 0.0001845,
|
|
"loss": 6.7495,
|
|
"mean_token_accuracy": 0.09524724259972572,
|
|
"num_tokens": 681758.0,
|
|
"step": 370
|
|
},
|
|
{
|
|
"entropy": 7.373961687088013,
|
|
"epoch": 0.3138075313807531,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.000187,
|
|
"loss": 6.7705,
|
|
"mean_token_accuracy": 0.0979244202375412,
|
|
"num_tokens": 690727.0,
|
|
"step": 375
|
|
},
|
|
{
|
|
"entropy": 7.338280773162841,
|
|
"epoch": 0.3179916317991632,
|
|
"grad_norm": 1.0703125,
|
|
"learning_rate": 0.0001895,
|
|
"loss": 6.7246,
|
|
"mean_token_accuracy": 0.09411645382642746,
|
|
"num_tokens": 701210.0,
|
|
"step": 380
|
|
},
|
|
{
|
|
"entropy": 7.166968154907226,
|
|
"epoch": 0.32217573221757323,
|
|
"grad_norm": 1.375,
|
|
"learning_rate": 0.000192,
|
|
"loss": 6.7198,
|
|
"mean_token_accuracy": 0.0960633248090744,
|
|
"num_tokens": 710869.0,
|
|
"step": 385
|
|
},
|
|
{
|
|
"entropy": 7.326888370513916,
|
|
"epoch": 0.3263598326359833,
|
|
"grad_norm": 1.0234375,
|
|
"learning_rate": 0.0001945,
|
|
"loss": 6.8285,
|
|
"mean_token_accuracy": 0.09366477653384209,
|
|
"num_tokens": 721455.0,
|
|
"step": 390
|
|
},
|
|
{
|
|
"entropy": 7.2980750560760494,
|
|
"epoch": 0.3305439330543933,
|
|
"grad_norm": 0.91015625,
|
|
"learning_rate": 0.00019700000000000002,
|
|
"loss": 6.7142,
|
|
"mean_token_accuracy": 0.10093756765127182,
|
|
"num_tokens": 730917.0,
|
|
"step": 395
|
|
},
|
|
{
|
|
"entropy": 7.274618291854859,
|
|
"epoch": 0.33472803347280333,
|
|
"grad_norm": 1.09375,
|
|
"learning_rate": 0.00019950000000000002,
|
|
"loss": 6.7402,
|
|
"mean_token_accuracy": 0.09866241216659546,
|
|
"num_tokens": 739302.0,
|
|
"step": 400
|
|
},
|
|
{
|
|
"entropy": 7.29223780632019,
|
|
"epoch": 0.3389121338912134,
|
|
"grad_norm": 1.1640625,
|
|
"learning_rate": 0.000202,
|
|
"loss": 6.7097,
|
|
"mean_token_accuracy": 0.09633656591176987,
|
|
"num_tokens": 748785.0,
|
|
"step": 405
|
|
},
|
|
{
|
|
"entropy": 7.214037179946899,
|
|
"epoch": 0.34309623430962344,
|
|
"grad_norm": 1.125,
|
|
"learning_rate": 0.00020449999999999998,
|
|
"loss": 6.6545,
|
|
"mean_token_accuracy": 0.10300204828381539,
|
|
"num_tokens": 757215.0,
|
|
"step": 410
|
|
},
|
|
{
|
|
"entropy": 7.234053421020508,
|
|
"epoch": 0.3472803347280335,
|
|
"grad_norm": 1.0859375,
|
|
"learning_rate": 0.000207,
|
|
"loss": 6.5793,
|
|
"mean_token_accuracy": 0.10455078259110451,
|
|
"num_tokens": 765682.0,
|
|
"step": 415
|
|
},
|
|
{
|
|
"entropy": 7.289216184616089,
|
|
"epoch": 0.3514644351464435,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 0.0002095,
|
|
"loss": 6.7796,
|
|
"mean_token_accuracy": 0.0947409562766552,
|
|
"num_tokens": 774814.0,
|
|
"step": 420
|
|
},
|
|
{
|
|
"entropy": 7.198604440689087,
|
|
"epoch": 0.35564853556485354,
|
|
"grad_norm": 1.2734375,
|
|
"learning_rate": 0.000212,
|
|
"loss": 6.6991,
|
|
"mean_token_accuracy": 0.1013828955590725,
|
|
"num_tokens": 784019.0,
|
|
"step": 425
|
|
},
|
|
{
|
|
"entropy": 7.1404036521911625,
|
|
"epoch": 0.3598326359832636,
|
|
"grad_norm": 1.03125,
|
|
"learning_rate": 0.0002145,
|
|
"loss": 6.5758,
|
|
"mean_token_accuracy": 0.10198531225323677,
|
|
"num_tokens": 793237.0,
|
|
"step": 430
|
|
},
|
|
{
|
|
"entropy": 7.1974271774292,
|
|
"epoch": 0.36401673640167365,
|
|
"grad_norm": 1.109375,
|
|
"learning_rate": 0.00021700000000000002,
|
|
"loss": 6.6162,
|
|
"mean_token_accuracy": 0.10484573319554329,
|
|
"num_tokens": 801548.0,
|
|
"step": 435
|
|
},
|
|
{
|
|
"entropy": 7.14229302406311,
|
|
"epoch": 0.3682008368200837,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.0002195,
|
|
"loss": 6.6647,
|
|
"mean_token_accuracy": 0.10372263044118882,
|
|
"num_tokens": 810974.0,
|
|
"step": 440
|
|
},
|
|
{
|
|
"entropy": 7.127859115600586,
|
|
"epoch": 0.3723849372384937,
|
|
"grad_norm": 1.1484375,
|
|
"learning_rate": 0.000222,
|
|
"loss": 6.5601,
|
|
"mean_token_accuracy": 0.1046792708337307,
|
|
"num_tokens": 820023.0,
|
|
"step": 445
|
|
},
|
|
{
|
|
"entropy": 7.167815637588501,
|
|
"epoch": 0.37656903765690375,
|
|
"grad_norm": 1.3046875,
|
|
"learning_rate": 0.0002245,
|
|
"loss": 6.644,
|
|
"mean_token_accuracy": 0.10097603350877762,
|
|
"num_tokens": 828297.0,
|
|
"step": 450
|
|
},
|
|
{
|
|
"entropy": 7.183742904663086,
|
|
"epoch": 0.3807531380753138,
|
|
"grad_norm": 1.265625,
|
|
"learning_rate": 0.00022700000000000002,
|
|
"loss": 6.693,
|
|
"mean_token_accuracy": 0.0922306977212429,
|
|
"num_tokens": 838268.0,
|
|
"step": 455
|
|
},
|
|
{
|
|
"entropy": 7.183096885681152,
|
|
"epoch": 0.38493723849372385,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 0.00022950000000000002,
|
|
"loss": 6.6117,
|
|
"mean_token_accuracy": 0.10094281509518624,
|
|
"num_tokens": 846491.0,
|
|
"step": 460
|
|
},
|
|
{
|
|
"entropy": 7.094890165328979,
|
|
"epoch": 0.3891213389121339,
|
|
"grad_norm": 1.046875,
|
|
"learning_rate": 0.00023200000000000003,
|
|
"loss": 6.5546,
|
|
"mean_token_accuracy": 0.10523919835686683,
|
|
"num_tokens": 856501.0,
|
|
"step": 465
|
|
},
|
|
{
|
|
"entropy": 7.092893457412719,
|
|
"epoch": 0.39330543933054396,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 0.00023449999999999998,
|
|
"loss": 6.5958,
|
|
"mean_token_accuracy": 0.10569338127970695,
|
|
"num_tokens": 867054.0,
|
|
"step": 470
|
|
},
|
|
{
|
|
"entropy": 7.0160187721252445,
|
|
"epoch": 0.39748953974895396,
|
|
"grad_norm": 1.0859375,
|
|
"learning_rate": 0.000237,
|
|
"loss": 6.6239,
|
|
"mean_token_accuracy": 0.10482471361756325,
|
|
"num_tokens": 875956.0,
|
|
"step": 475
|
|
},
|
|
{
|
|
"entropy": 7.013614892959595,
|
|
"epoch": 0.401673640167364,
|
|
"grad_norm": 1.3671875,
|
|
"learning_rate": 0.0002395,
|
|
"loss": 6.5334,
|
|
"mean_token_accuracy": 0.1062661923468113,
|
|
"num_tokens": 884427.0,
|
|
"step": 480
|
|
},
|
|
{
|
|
"entropy": 7.170659828186035,
|
|
"epoch": 0.40585774058577406,
|
|
"grad_norm": 0.9765625,
|
|
"learning_rate": 0.000242,
|
|
"loss": 6.6249,
|
|
"mean_token_accuracy": 0.10283279418945312,
|
|
"num_tokens": 894181.0,
|
|
"step": 485
|
|
},
|
|
{
|
|
"entropy": 6.948815584182739,
|
|
"epoch": 0.4100418410041841,
|
|
"grad_norm": 1.0703125,
|
|
"learning_rate": 0.0002445,
|
|
"loss": 6.4707,
|
|
"mean_token_accuracy": 0.10783380344510078,
|
|
"num_tokens": 904076.0,
|
|
"step": 490
|
|
},
|
|
{
|
|
"entropy": 7.047258281707764,
|
|
"epoch": 0.41422594142259417,
|
|
"grad_norm": 0.953125,
|
|
"learning_rate": 0.000247,
|
|
"loss": 6.6131,
|
|
"mean_token_accuracy": 0.10789759010076523,
|
|
"num_tokens": 913915.0,
|
|
"step": 495
|
|
},
|
|
{
|
|
"entropy": 7.123381805419922,
|
|
"epoch": 0.41841004184100417,
|
|
"grad_norm": 1.1640625,
|
|
"learning_rate": 0.0002495,
|
|
"loss": 6.6688,
|
|
"mean_token_accuracy": 0.1050674706697464,
|
|
"num_tokens": 922876.0,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.41841004184100417,
|
|
"eval_entropy": 7.012504460621465,
|
|
"eval_loss": 6.6101908683776855,
|
|
"eval_mean_token_accuracy": 0.10924081287226795,
|
|
"eval_num_tokens": 922876.0,
|
|
"eval_runtime": 2.0948,
|
|
"eval_samples_per_second": 1865.089,
|
|
"eval_steps_per_second": 233.434,
|
|
"step": 500
|
|
}
|
|
],
|
|
"logging_steps": 5,
|
|
"max_steps": 11950,
|
|
"num_input_tokens_seen": 0,
|
|
"num_train_epochs": 10,
|
|
"save_steps": 500,
|
|
"stateful_callbacks": {
|
|
"TrainerControl": {
|
|
"args": {
|
|
"should_epoch_stop": false,
|
|
"should_evaluate": false,
|
|
"should_log": false,
|
|
"should_save": true,
|
|
"should_training_stop": false
|
|
},
|
|
"attributes": {}
|
|
}
|
|
},
|
|
"total_flos": 200351734824960.0,
|
|
"train_batch_size": 16,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|