1046 lines
28 KiB
JSON
1046 lines
28 KiB
JSON
{
|
|
"best_global_step": null,
|
|
"best_metric": null,
|
|
"best_model_checkpoint": null,
|
|
"epoch": 0.4142502071251036,
|
|
"eval_steps": 500,
|
|
"global_step": 500,
|
|
"is_hyper_param_search": false,
|
|
"is_local_process_zero": true,
|
|
"is_world_process_zero": true,
|
|
"log_history": [
|
|
{
|
|
"entropy": 10.691956996917725,
|
|
"epoch": 0.004142502071251036,
|
|
"grad_norm": 12.6875,
|
|
"learning_rate": 2e-06,
|
|
"loss": 10.786,
|
|
"mean_token_accuracy": 0.0,
|
|
"num_tokens": 8553.0,
|
|
"step": 5
|
|
},
|
|
{
|
|
"entropy": 10.69189453125,
|
|
"epoch": 0.008285004142502071,
|
|
"grad_norm": 13.0625,
|
|
"learning_rate": 4.5e-06,
|
|
"loss": 10.7413,
|
|
"mean_token_accuracy": 0.00027811190811917187,
|
|
"num_tokens": 18133.0,
|
|
"step": 10
|
|
},
|
|
{
|
|
"entropy": 10.691031455993652,
|
|
"epoch": 0.012427506213753107,
|
|
"grad_norm": 9.5625,
|
|
"learning_rate": 7e-06,
|
|
"loss": 10.4928,
|
|
"mean_token_accuracy": 0.023286720411852,
|
|
"num_tokens": 27152.0,
|
|
"step": 15
|
|
},
|
|
{
|
|
"entropy": 10.681630420684815,
|
|
"epoch": 0.016570008285004142,
|
|
"grad_norm": 6.40625,
|
|
"learning_rate": 9.5e-06,
|
|
"loss": 10.2019,
|
|
"mean_token_accuracy": 0.042532872036099435,
|
|
"num_tokens": 35844.0,
|
|
"step": 20
|
|
},
|
|
{
|
|
"entropy": 10.63299388885498,
|
|
"epoch": 0.020712510356255178,
|
|
"grad_norm": 4.46875,
|
|
"learning_rate": 1.2e-05,
|
|
"loss": 9.8879,
|
|
"mean_token_accuracy": 0.04524078518152237,
|
|
"num_tokens": 44730.0,
|
|
"step": 25
|
|
},
|
|
{
|
|
"entropy": 10.594239139556885,
|
|
"epoch": 0.024855012427506214,
|
|
"grad_norm": 3.6875,
|
|
"learning_rate": 1.4500000000000002e-05,
|
|
"loss": 9.7944,
|
|
"mean_token_accuracy": 0.04270199164748192,
|
|
"num_tokens": 53751.0,
|
|
"step": 30
|
|
},
|
|
{
|
|
"entropy": 10.599893474578858,
|
|
"epoch": 0.02899751449875725,
|
|
"grad_norm": 2.9375,
|
|
"learning_rate": 1.7000000000000003e-05,
|
|
"loss": 9.6813,
|
|
"mean_token_accuracy": 0.0431942991912365,
|
|
"num_tokens": 62399.0,
|
|
"step": 35
|
|
},
|
|
{
|
|
"entropy": 10.593089294433593,
|
|
"epoch": 0.033140016570008285,
|
|
"grad_norm": 2.734375,
|
|
"learning_rate": 1.95e-05,
|
|
"loss": 9.675,
|
|
"mean_token_accuracy": 0.045351412147283554,
|
|
"num_tokens": 72384.0,
|
|
"step": 40
|
|
},
|
|
{
|
|
"entropy": 10.579291725158692,
|
|
"epoch": 0.037282518641259324,
|
|
"grad_norm": 2.71875,
|
|
"learning_rate": 2.2e-05,
|
|
"loss": 9.625,
|
|
"mean_token_accuracy": 0.03919882737100124,
|
|
"num_tokens": 81720.0,
|
|
"step": 45
|
|
},
|
|
{
|
|
"entropy": 10.570260047912598,
|
|
"epoch": 0.041425020712510356,
|
|
"grad_norm": 2.34375,
|
|
"learning_rate": 2.4500000000000003e-05,
|
|
"loss": 9.5545,
|
|
"mean_token_accuracy": 0.045346799120306966,
|
|
"num_tokens": 91634.0,
|
|
"step": 50
|
|
},
|
|
{
|
|
"entropy": 10.559380340576173,
|
|
"epoch": 0.045567522783761395,
|
|
"grad_norm": 2.28125,
|
|
"learning_rate": 2.7e-05,
|
|
"loss": 9.4963,
|
|
"mean_token_accuracy": 0.048221738263964656,
|
|
"num_tokens": 100359.0,
|
|
"step": 55
|
|
},
|
|
{
|
|
"entropy": 10.541158962249757,
|
|
"epoch": 0.04971002485501243,
|
|
"grad_norm": 2.390625,
|
|
"learning_rate": 2.95e-05,
|
|
"loss": 9.4361,
|
|
"mean_token_accuracy": 0.05027815327048302,
|
|
"num_tokens": 108678.0,
|
|
"step": 60
|
|
},
|
|
{
|
|
"entropy": 10.408777332305908,
|
|
"epoch": 0.053852526926263466,
|
|
"grad_norm": 2.359375,
|
|
"learning_rate": 3.2e-05,
|
|
"loss": 9.3154,
|
|
"mean_token_accuracy": 0.055302799120545386,
|
|
"num_tokens": 117022.0,
|
|
"step": 65
|
|
},
|
|
{
|
|
"entropy": 10.43354835510254,
|
|
"epoch": 0.0579950289975145,
|
|
"grad_norm": 2.234375,
|
|
"learning_rate": 3.4500000000000005e-05,
|
|
"loss": 9.2758,
|
|
"mean_token_accuracy": 0.05044409707188606,
|
|
"num_tokens": 126489.0,
|
|
"step": 70
|
|
},
|
|
{
|
|
"entropy": 10.479030323028564,
|
|
"epoch": 0.06213753106876554,
|
|
"grad_norm": 2.265625,
|
|
"learning_rate": 3.7e-05,
|
|
"loss": 9.1042,
|
|
"mean_token_accuracy": 0.060464993864297864,
|
|
"num_tokens": 135926.0,
|
|
"step": 75
|
|
},
|
|
{
|
|
"entropy": 10.352683162689209,
|
|
"epoch": 0.06628003314001657,
|
|
"grad_norm": 2.1875,
|
|
"learning_rate": 3.95e-05,
|
|
"loss": 9.0851,
|
|
"mean_token_accuracy": 0.05333640910685063,
|
|
"num_tokens": 144488.0,
|
|
"step": 80
|
|
},
|
|
{
|
|
"entropy": 10.436153030395507,
|
|
"epoch": 0.07042253521126761,
|
|
"grad_norm": 2.171875,
|
|
"learning_rate": 4.2000000000000004e-05,
|
|
"loss": 8.9888,
|
|
"mean_token_accuracy": 0.05590510331094265,
|
|
"num_tokens": 154180.0,
|
|
"step": 85
|
|
},
|
|
{
|
|
"entropy": 10.357822036743164,
|
|
"epoch": 0.07456503728251865,
|
|
"grad_norm": 2.28125,
|
|
"learning_rate": 4.45e-05,
|
|
"loss": 8.9305,
|
|
"mean_token_accuracy": 0.05672054700553417,
|
|
"num_tokens": 163448.0,
|
|
"step": 90
|
|
},
|
|
{
|
|
"entropy": 10.390939235687256,
|
|
"epoch": 0.07870753935376967,
|
|
"grad_norm": 2.140625,
|
|
"learning_rate": 4.7000000000000004e-05,
|
|
"loss": 8.8567,
|
|
"mean_token_accuracy": 0.054386957362294194,
|
|
"num_tokens": 173318.0,
|
|
"step": 95
|
|
},
|
|
{
|
|
"entropy": 10.244186305999756,
|
|
"epoch": 0.08285004142502071,
|
|
"grad_norm": 2.046875,
|
|
"learning_rate": 4.9500000000000004e-05,
|
|
"loss": 8.745,
|
|
"mean_token_accuracy": 0.05825656205415726,
|
|
"num_tokens": 181540.0,
|
|
"step": 100
|
|
},
|
|
{
|
|
"entropy": 10.197363090515136,
|
|
"epoch": 0.08699254349627175,
|
|
"grad_norm": 1.9609375,
|
|
"learning_rate": 5.2e-05,
|
|
"loss": 8.5793,
|
|
"mean_token_accuracy": 0.05952193439006805,
|
|
"num_tokens": 190932.0,
|
|
"step": 105
|
|
},
|
|
{
|
|
"entropy": 10.11266918182373,
|
|
"epoch": 0.09113504556752279,
|
|
"grad_norm": 2.09375,
|
|
"learning_rate": 5.45e-05,
|
|
"loss": 8.4206,
|
|
"mean_token_accuracy": 0.0623247817158699,
|
|
"num_tokens": 199793.0,
|
|
"step": 110
|
|
},
|
|
{
|
|
"entropy": 10.071755409240723,
|
|
"epoch": 0.09527754763877382,
|
|
"grad_norm": 2.0,
|
|
"learning_rate": 5.7e-05,
|
|
"loss": 8.4016,
|
|
"mean_token_accuracy": 0.056386180222034454,
|
|
"num_tokens": 208865.0,
|
|
"step": 115
|
|
},
|
|
{
|
|
"entropy": 9.93977689743042,
|
|
"epoch": 0.09942004971002485,
|
|
"grad_norm": 1.7578125,
|
|
"learning_rate": 5.9499999999999996e-05,
|
|
"loss": 8.2482,
|
|
"mean_token_accuracy": 0.06587636284530163,
|
|
"num_tokens": 218987.0,
|
|
"step": 120
|
|
},
|
|
{
|
|
"entropy": 9.830070209503173,
|
|
"epoch": 0.1035625517812759,
|
|
"grad_norm": 1.6171875,
|
|
"learning_rate": 6.2e-05,
|
|
"loss": 8.1893,
|
|
"mean_token_accuracy": 0.06974356137216091,
|
|
"num_tokens": 228815.0,
|
|
"step": 125
|
|
},
|
|
{
|
|
"entropy": 9.623151206970215,
|
|
"epoch": 0.10770505385252693,
|
|
"grad_norm": 1.5546875,
|
|
"learning_rate": 6.450000000000001e-05,
|
|
"loss": 8.0572,
|
|
"mean_token_accuracy": 0.06595751233398914,
|
|
"num_tokens": 237752.0,
|
|
"step": 130
|
|
},
|
|
{
|
|
"entropy": 9.47017993927002,
|
|
"epoch": 0.11184755592377796,
|
|
"grad_norm": 1.6171875,
|
|
"learning_rate": 6.7e-05,
|
|
"loss": 7.9063,
|
|
"mean_token_accuracy": 0.07446046024560929,
|
|
"num_tokens": 246512.0,
|
|
"step": 135
|
|
},
|
|
{
|
|
"entropy": 9.312911701202392,
|
|
"epoch": 0.115990057995029,
|
|
"grad_norm": 1.8359375,
|
|
"learning_rate": 6.950000000000001e-05,
|
|
"loss": 7.8915,
|
|
"mean_token_accuracy": 0.0656554326415062,
|
|
"num_tokens": 255300.0,
|
|
"step": 140
|
|
},
|
|
{
|
|
"entropy": 9.113032817840576,
|
|
"epoch": 0.12013256006628004,
|
|
"grad_norm": 1.4609375,
|
|
"learning_rate": 7.2e-05,
|
|
"loss": 7.7471,
|
|
"mean_token_accuracy": 0.06687054596841335,
|
|
"num_tokens": 265125.0,
|
|
"step": 145
|
|
},
|
|
{
|
|
"entropy": 8.8920316696167,
|
|
"epoch": 0.12427506213753108,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 7.45e-05,
|
|
"loss": 7.6598,
|
|
"mean_token_accuracy": 0.06876451000571251,
|
|
"num_tokens": 274516.0,
|
|
"step": 150
|
|
},
|
|
{
|
|
"entropy": 8.746868896484376,
|
|
"epoch": 0.12841756420878211,
|
|
"grad_norm": 1.2265625,
|
|
"learning_rate": 7.7e-05,
|
|
"loss": 7.6736,
|
|
"mean_token_accuracy": 0.07511950880289078,
|
|
"num_tokens": 282994.0,
|
|
"step": 155
|
|
},
|
|
{
|
|
"entropy": 8.574301624298096,
|
|
"epoch": 0.13256006628003314,
|
|
"grad_norm": 1.3984375,
|
|
"learning_rate": 7.950000000000001e-05,
|
|
"loss": 7.5639,
|
|
"mean_token_accuracy": 0.07003462612628937,
|
|
"num_tokens": 291703.0,
|
|
"step": 160
|
|
},
|
|
{
|
|
"entropy": 8.47190055847168,
|
|
"epoch": 0.13670256835128416,
|
|
"grad_norm": 1.203125,
|
|
"learning_rate": 8.2e-05,
|
|
"loss": 7.4587,
|
|
"mean_token_accuracy": 0.0722161963582039,
|
|
"num_tokens": 301596.0,
|
|
"step": 165
|
|
},
|
|
{
|
|
"entropy": 8.276020717620849,
|
|
"epoch": 0.14084507042253522,
|
|
"grad_norm": 1.3515625,
|
|
"learning_rate": 8.450000000000001e-05,
|
|
"loss": 7.5039,
|
|
"mean_token_accuracy": 0.075354777649045,
|
|
"num_tokens": 309678.0,
|
|
"step": 170
|
|
},
|
|
{
|
|
"entropy": 8.170723915100098,
|
|
"epoch": 0.14498757249378624,
|
|
"grad_norm": 1.390625,
|
|
"learning_rate": 8.7e-05,
|
|
"loss": 7.4176,
|
|
"mean_token_accuracy": 0.07923398241400718,
|
|
"num_tokens": 319046.0,
|
|
"step": 175
|
|
},
|
|
{
|
|
"entropy": 8.180229568481446,
|
|
"epoch": 0.1491300745650373,
|
|
"grad_norm": 1.453125,
|
|
"learning_rate": 8.95e-05,
|
|
"loss": 7.3691,
|
|
"mean_token_accuracy": 0.07866172268986701,
|
|
"num_tokens": 328004.0,
|
|
"step": 180
|
|
},
|
|
{
|
|
"entropy": 8.078632307052612,
|
|
"epoch": 0.15327257663628832,
|
|
"grad_norm": 1.3203125,
|
|
"learning_rate": 9.2e-05,
|
|
"loss": 7.3784,
|
|
"mean_token_accuracy": 0.0761629119515419,
|
|
"num_tokens": 337455.0,
|
|
"step": 185
|
|
},
|
|
{
|
|
"entropy": 8.00011487007141,
|
|
"epoch": 0.15741507870753935,
|
|
"grad_norm": 1.5078125,
|
|
"learning_rate": 9.45e-05,
|
|
"loss": 7.3504,
|
|
"mean_token_accuracy": 0.07819610238075256,
|
|
"num_tokens": 345631.0,
|
|
"step": 190
|
|
},
|
|
{
|
|
"entropy": 7.865444993972778,
|
|
"epoch": 0.1615575807787904,
|
|
"grad_norm": 1.421875,
|
|
"learning_rate": 9.7e-05,
|
|
"loss": 7.2621,
|
|
"mean_token_accuracy": 0.07181993611156941,
|
|
"num_tokens": 354790.0,
|
|
"step": 195
|
|
},
|
|
{
|
|
"entropy": 7.92261233329773,
|
|
"epoch": 0.16570008285004142,
|
|
"grad_norm": 1.46875,
|
|
"learning_rate": 9.95e-05,
|
|
"loss": 7.374,
|
|
"mean_token_accuracy": 0.08448374196887017,
|
|
"num_tokens": 364480.0,
|
|
"step": 200
|
|
},
|
|
{
|
|
"entropy": 7.90945405960083,
|
|
"epoch": 0.16984258492129245,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 0.000102,
|
|
"loss": 7.4125,
|
|
"mean_token_accuracy": 0.07339793667197228,
|
|
"num_tokens": 372751.0,
|
|
"step": 205
|
|
},
|
|
{
|
|
"entropy": 7.968420600891113,
|
|
"epoch": 0.1739850869925435,
|
|
"grad_norm": 1.59375,
|
|
"learning_rate": 0.00010449999999999999,
|
|
"loss": 7.3969,
|
|
"mean_token_accuracy": 0.07574166990816593,
|
|
"num_tokens": 381140.0,
|
|
"step": 210
|
|
},
|
|
{
|
|
"entropy": 7.890913581848144,
|
|
"epoch": 0.17812758906379453,
|
|
"grad_norm": 1.421875,
|
|
"learning_rate": 0.000107,
|
|
"loss": 7.3887,
|
|
"mean_token_accuracy": 0.07703622430562973,
|
|
"num_tokens": 391415.0,
|
|
"step": 215
|
|
},
|
|
{
|
|
"entropy": 7.813537740707398,
|
|
"epoch": 0.18227009113504558,
|
|
"grad_norm": 1.1484375,
|
|
"learning_rate": 0.0001095,
|
|
"loss": 7.3673,
|
|
"mean_token_accuracy": 0.07801026552915573,
|
|
"num_tokens": 400174.0,
|
|
"step": 220
|
|
},
|
|
{
|
|
"entropy": 7.883461046218872,
|
|
"epoch": 0.1864125932062966,
|
|
"grad_norm": 1.5390625,
|
|
"learning_rate": 0.000112,
|
|
"loss": 7.2726,
|
|
"mean_token_accuracy": 0.08017019852995873,
|
|
"num_tokens": 409355.0,
|
|
"step": 225
|
|
},
|
|
{
|
|
"entropy": 7.8811296939849855,
|
|
"epoch": 0.19055509527754763,
|
|
"grad_norm": 1.3515625,
|
|
"learning_rate": 0.0001145,
|
|
"loss": 7.2959,
|
|
"mean_token_accuracy": 0.077110555768013,
|
|
"num_tokens": 418336.0,
|
|
"step": 230
|
|
},
|
|
{
|
|
"entropy": 7.754353618621826,
|
|
"epoch": 0.19469759734879868,
|
|
"grad_norm": 1.375,
|
|
"learning_rate": 0.00011700000000000001,
|
|
"loss": 7.292,
|
|
"mean_token_accuracy": 0.0773789893835783,
|
|
"num_tokens": 427771.0,
|
|
"step": 235
|
|
},
|
|
{
|
|
"entropy": 7.748405838012696,
|
|
"epoch": 0.1988400994200497,
|
|
"grad_norm": 1.5859375,
|
|
"learning_rate": 0.00011949999999999999,
|
|
"loss": 7.3225,
|
|
"mean_token_accuracy": 0.07635865435004234,
|
|
"num_tokens": 435576.0,
|
|
"step": 240
|
|
},
|
|
{
|
|
"entropy": 7.725781488418579,
|
|
"epoch": 0.20298260149130073,
|
|
"grad_norm": 1.5546875,
|
|
"learning_rate": 0.000122,
|
|
"loss": 7.2366,
|
|
"mean_token_accuracy": 0.07985471300780773,
|
|
"num_tokens": 444360.0,
|
|
"step": 245
|
|
},
|
|
{
|
|
"entropy": 7.684485578536988,
|
|
"epoch": 0.2071251035625518,
|
|
"grad_norm": 1.1328125,
|
|
"learning_rate": 0.0001245,
|
|
"loss": 7.1657,
|
|
"mean_token_accuracy": 0.08312897011637688,
|
|
"num_tokens": 452946.0,
|
|
"step": 250
|
|
},
|
|
{
|
|
"entropy": 7.705522155761718,
|
|
"epoch": 0.2112676056338028,
|
|
"grad_norm": 1.2109375,
|
|
"learning_rate": 0.000127,
|
|
"loss": 7.1751,
|
|
"mean_token_accuracy": 0.08343161344528198,
|
|
"num_tokens": 462105.0,
|
|
"step": 255
|
|
},
|
|
{
|
|
"entropy": 7.676226043701172,
|
|
"epoch": 0.21541010770505387,
|
|
"grad_norm": 1.2421875,
|
|
"learning_rate": 0.0001295,
|
|
"loss": 7.3555,
|
|
"mean_token_accuracy": 0.07977722287178039,
|
|
"num_tokens": 472397.0,
|
|
"step": 260
|
|
},
|
|
{
|
|
"entropy": 7.740552568435669,
|
|
"epoch": 0.2195526097763049,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 0.000132,
|
|
"loss": 7.2001,
|
|
"mean_token_accuracy": 0.078244648873806,
|
|
"num_tokens": 481595.0,
|
|
"step": 265
|
|
},
|
|
{
|
|
"entropy": 7.51032190322876,
|
|
"epoch": 0.22369511184755592,
|
|
"grad_norm": 1.2265625,
|
|
"learning_rate": 0.00013450000000000002,
|
|
"loss": 7.0929,
|
|
"mean_token_accuracy": 0.08391840234398842,
|
|
"num_tokens": 490034.0,
|
|
"step": 270
|
|
},
|
|
{
|
|
"entropy": 7.6380149841308596,
|
|
"epoch": 0.22783761391880697,
|
|
"grad_norm": 1.5390625,
|
|
"learning_rate": 0.00013700000000000002,
|
|
"loss": 7.2089,
|
|
"mean_token_accuracy": 0.08496099263429642,
|
|
"num_tokens": 498521.0,
|
|
"step": 275
|
|
},
|
|
{
|
|
"entropy": 7.588693332672119,
|
|
"epoch": 0.231980115990058,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 0.0001395,
|
|
"loss": 7.1605,
|
|
"mean_token_accuracy": 0.07962968274950981,
|
|
"num_tokens": 507251.0,
|
|
"step": 280
|
|
},
|
|
{
|
|
"entropy": 7.5256298065185545,
|
|
"epoch": 0.23612261806130902,
|
|
"grad_norm": 1.2265625,
|
|
"learning_rate": 0.00014199999999999998,
|
|
"loss": 7.0795,
|
|
"mean_token_accuracy": 0.08021439425647259,
|
|
"num_tokens": 516507.0,
|
|
"step": 285
|
|
},
|
|
{
|
|
"entropy": 7.683542346954345,
|
|
"epoch": 0.24026512013256007,
|
|
"grad_norm": 1.421875,
|
|
"learning_rate": 0.0001445,
|
|
"loss": 7.1987,
|
|
"mean_token_accuracy": 0.07887389063835144,
|
|
"num_tokens": 526322.0,
|
|
"step": 290
|
|
},
|
|
{
|
|
"entropy": 7.589657783508301,
|
|
"epoch": 0.2444076222038111,
|
|
"grad_norm": 1.3671875,
|
|
"learning_rate": 0.000147,
|
|
"loss": 7.0698,
|
|
"mean_token_accuracy": 0.0887388363480568,
|
|
"num_tokens": 534762.0,
|
|
"step": 295
|
|
},
|
|
{
|
|
"entropy": 7.573185110092163,
|
|
"epoch": 0.24855012427506215,
|
|
"grad_norm": 1.1953125,
|
|
"learning_rate": 0.0001495,
|
|
"loss": 7.128,
|
|
"mean_token_accuracy": 0.08384166359901428,
|
|
"num_tokens": 544279.0,
|
|
"step": 300
|
|
},
|
|
{
|
|
"entropy": 7.464304256439209,
|
|
"epoch": 0.2526926263463132,
|
|
"grad_norm": 1.1171875,
|
|
"learning_rate": 0.000152,
|
|
"loss": 7.1072,
|
|
"mean_token_accuracy": 0.08207703456282615,
|
|
"num_tokens": 553178.0,
|
|
"step": 305
|
|
},
|
|
{
|
|
"entropy": 7.532624387741089,
|
|
"epoch": 0.25683512841756423,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 0.00015450000000000001,
|
|
"loss": 7.0674,
|
|
"mean_token_accuracy": 0.08325572535395623,
|
|
"num_tokens": 562383.0,
|
|
"step": 310
|
|
},
|
|
{
|
|
"entropy": 7.475009918212891,
|
|
"epoch": 0.2609776304888152,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 0.000157,
|
|
"loss": 7.0478,
|
|
"mean_token_accuracy": 0.08197091519832611,
|
|
"num_tokens": 571104.0,
|
|
"step": 315
|
|
},
|
|
{
|
|
"entropy": 7.402563810348511,
|
|
"epoch": 0.2651201325600663,
|
|
"grad_norm": 1.2421875,
|
|
"learning_rate": 0.0001595,
|
|
"loss": 6.9822,
|
|
"mean_token_accuracy": 0.08756032511591912,
|
|
"num_tokens": 579890.0,
|
|
"step": 320
|
|
},
|
|
{
|
|
"entropy": 7.563023519515991,
|
|
"epoch": 0.26926263463131733,
|
|
"grad_norm": 1.203125,
|
|
"learning_rate": 0.000162,
|
|
"loss": 7.1609,
|
|
"mean_token_accuracy": 0.08216102197766303,
|
|
"num_tokens": 589033.0,
|
|
"step": 325
|
|
},
|
|
{
|
|
"entropy": 7.385844802856445,
|
|
"epoch": 0.27340513670256833,
|
|
"grad_norm": 1.1953125,
|
|
"learning_rate": 0.00016450000000000001,
|
|
"loss": 7.031,
|
|
"mean_token_accuracy": 0.08240275979042053,
|
|
"num_tokens": 598040.0,
|
|
"step": 330
|
|
},
|
|
{
|
|
"entropy": 7.4914178371429445,
|
|
"epoch": 0.2775476387738194,
|
|
"grad_norm": 1.1640625,
|
|
"learning_rate": 0.00016700000000000002,
|
|
"loss": 7.1792,
|
|
"mean_token_accuracy": 0.08133822605013848,
|
|
"num_tokens": 607618.0,
|
|
"step": 335
|
|
},
|
|
{
|
|
"entropy": 7.360255098342895,
|
|
"epoch": 0.28169014084507044,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 0.00016950000000000003,
|
|
"loss": 6.9847,
|
|
"mean_token_accuracy": 0.08660589605569839,
|
|
"num_tokens": 616107.0,
|
|
"step": 340
|
|
},
|
|
{
|
|
"entropy": 7.469431829452515,
|
|
"epoch": 0.28583264291632143,
|
|
"grad_norm": 1.25,
|
|
"learning_rate": 0.00017199999999999998,
|
|
"loss": 7.0424,
|
|
"mean_token_accuracy": 0.08697322458028793,
|
|
"num_tokens": 625954.0,
|
|
"step": 345
|
|
},
|
|
{
|
|
"entropy": 7.43549690246582,
|
|
"epoch": 0.2899751449875725,
|
|
"grad_norm": 1.3203125,
|
|
"learning_rate": 0.00017449999999999999,
|
|
"loss": 7.1473,
|
|
"mean_token_accuracy": 0.07502747476100921,
|
|
"num_tokens": 635840.0,
|
|
"step": 350
|
|
},
|
|
{
|
|
"entropy": 7.531662464141846,
|
|
"epoch": 0.29411764705882354,
|
|
"grad_norm": 1.2265625,
|
|
"learning_rate": 0.000177,
|
|
"loss": 7.0981,
|
|
"mean_token_accuracy": 0.08479553908109665,
|
|
"num_tokens": 644597.0,
|
|
"step": 355
|
|
},
|
|
{
|
|
"entropy": 7.403447055816651,
|
|
"epoch": 0.2982601491300746,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 0.0001795,
|
|
"loss": 7.019,
|
|
"mean_token_accuracy": 0.0871436946094036,
|
|
"num_tokens": 653454.0,
|
|
"step": 360
|
|
},
|
|
{
|
|
"entropy": 7.331816291809082,
|
|
"epoch": 0.3024026512013256,
|
|
"grad_norm": 1.3828125,
|
|
"learning_rate": 0.000182,
|
|
"loss": 7.0644,
|
|
"mean_token_accuracy": 0.08571388535201549,
|
|
"num_tokens": 662524.0,
|
|
"step": 365
|
|
},
|
|
{
|
|
"entropy": 7.350220823287964,
|
|
"epoch": 0.30654515327257664,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 0.0001845,
|
|
"loss": 7.037,
|
|
"mean_token_accuracy": 0.08441569060087203,
|
|
"num_tokens": 672741.0,
|
|
"step": 370
|
|
},
|
|
{
|
|
"entropy": 7.462545442581177,
|
|
"epoch": 0.3106876553438277,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 0.000187,
|
|
"loss": 7.0014,
|
|
"mean_token_accuracy": 0.08310677483677864,
|
|
"num_tokens": 682068.0,
|
|
"step": 375
|
|
},
|
|
{
|
|
"entropy": 7.483937978744507,
|
|
"epoch": 0.3148301574150787,
|
|
"grad_norm": 1.140625,
|
|
"learning_rate": 0.0001895,
|
|
"loss": 7.068,
|
|
"mean_token_accuracy": 0.08450363203883171,
|
|
"num_tokens": 691506.0,
|
|
"step": 380
|
|
},
|
|
{
|
|
"entropy": 7.211332607269287,
|
|
"epoch": 0.31897265948632975,
|
|
"grad_norm": 1.375,
|
|
"learning_rate": 0.000192,
|
|
"loss": 6.9459,
|
|
"mean_token_accuracy": 0.07925326749682426,
|
|
"num_tokens": 702355.0,
|
|
"step": 385
|
|
},
|
|
{
|
|
"entropy": 7.426611804962159,
|
|
"epoch": 0.3231151615575808,
|
|
"grad_norm": 1.1796875,
|
|
"learning_rate": 0.0001945,
|
|
"loss": 7.0277,
|
|
"mean_token_accuracy": 0.08726275414228439,
|
|
"num_tokens": 710878.0,
|
|
"step": 390
|
|
},
|
|
{
|
|
"entropy": 7.349923038482666,
|
|
"epoch": 0.3272576636288318,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 0.00019700000000000002,
|
|
"loss": 7.0066,
|
|
"mean_token_accuracy": 0.0816049836575985,
|
|
"num_tokens": 719389.0,
|
|
"step": 395
|
|
},
|
|
{
|
|
"entropy": 7.291773509979248,
|
|
"epoch": 0.33140016570008285,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 0.00019950000000000002,
|
|
"loss": 6.9038,
|
|
"mean_token_accuracy": 0.08625322580337524,
|
|
"num_tokens": 728109.0,
|
|
"step": 400
|
|
},
|
|
{
|
|
"entropy": 7.3569948196411135,
|
|
"epoch": 0.3355426677713339,
|
|
"grad_norm": 1.3359375,
|
|
"learning_rate": 0.000202,
|
|
"loss": 7.0035,
|
|
"mean_token_accuracy": 0.08949297964572907,
|
|
"num_tokens": 737617.0,
|
|
"step": 405
|
|
},
|
|
{
|
|
"entropy": 7.196184492111206,
|
|
"epoch": 0.3396851698425849,
|
|
"grad_norm": 1.296875,
|
|
"learning_rate": 0.00020449999999999998,
|
|
"loss": 6.8985,
|
|
"mean_token_accuracy": 0.086442032456398,
|
|
"num_tokens": 746368.0,
|
|
"step": 410
|
|
},
|
|
{
|
|
"entropy": 7.21246075630188,
|
|
"epoch": 0.34382767191383595,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 0.000207,
|
|
"loss": 6.8542,
|
|
"mean_token_accuracy": 0.09095343127846718,
|
|
"num_tokens": 755561.0,
|
|
"step": 415
|
|
},
|
|
{
|
|
"entropy": 7.358605766296387,
|
|
"epoch": 0.347970173985087,
|
|
"grad_norm": 1.2578125,
|
|
"learning_rate": 0.0002095,
|
|
"loss": 6.9435,
|
|
"mean_token_accuracy": 0.08809717670083046,
|
|
"num_tokens": 764501.0,
|
|
"step": 420
|
|
},
|
|
{
|
|
"entropy": 7.229580879211426,
|
|
"epoch": 0.352112676056338,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 0.000212,
|
|
"loss": 7.085,
|
|
"mean_token_accuracy": 0.08461654633283615,
|
|
"num_tokens": 774003.0,
|
|
"step": 425
|
|
},
|
|
{
|
|
"entropy": 7.327147531509399,
|
|
"epoch": 0.35625517812758906,
|
|
"grad_norm": 1.15625,
|
|
"learning_rate": 0.0002145,
|
|
"loss": 6.8912,
|
|
"mean_token_accuracy": 0.09031773805618286,
|
|
"num_tokens": 782319.0,
|
|
"step": 430
|
|
},
|
|
{
|
|
"entropy": 7.148540925979614,
|
|
"epoch": 0.3603976801988401,
|
|
"grad_norm": 1.1328125,
|
|
"learning_rate": 0.00021700000000000002,
|
|
"loss": 6.8587,
|
|
"mean_token_accuracy": 0.08761301413178443,
|
|
"num_tokens": 791467.0,
|
|
"step": 435
|
|
},
|
|
{
|
|
"entropy": 7.317266750335693,
|
|
"epoch": 0.36454018227009116,
|
|
"grad_norm": 1.2109375,
|
|
"learning_rate": 0.0002195,
|
|
"loss": 6.9807,
|
|
"mean_token_accuracy": 0.08628286868333816,
|
|
"num_tokens": 800328.0,
|
|
"step": 440
|
|
},
|
|
{
|
|
"entropy": 7.229923439025879,
|
|
"epoch": 0.36868268434134216,
|
|
"grad_norm": 1.484375,
|
|
"learning_rate": 0.000222,
|
|
"loss": 6.9734,
|
|
"mean_token_accuracy": 0.08270182609558105,
|
|
"num_tokens": 808849.0,
|
|
"step": 445
|
|
},
|
|
{
|
|
"entropy": 7.28882212638855,
|
|
"epoch": 0.3728251864125932,
|
|
"grad_norm": 1.125,
|
|
"learning_rate": 0.0002245,
|
|
"loss": 6.9481,
|
|
"mean_token_accuracy": 0.09033941775560379,
|
|
"num_tokens": 819090.0,
|
|
"step": 450
|
|
},
|
|
{
|
|
"entropy": 7.23458833694458,
|
|
"epoch": 0.37696768848384427,
|
|
"grad_norm": 1.171875,
|
|
"learning_rate": 0.00022700000000000002,
|
|
"loss": 6.9226,
|
|
"mean_token_accuracy": 0.08801109045743942,
|
|
"num_tokens": 828527.0,
|
|
"step": 455
|
|
},
|
|
{
|
|
"entropy": 7.252940464019775,
|
|
"epoch": 0.38111019055509526,
|
|
"grad_norm": 1.1875,
|
|
"learning_rate": 0.00022950000000000002,
|
|
"loss": 6.9447,
|
|
"mean_token_accuracy": 0.08754967637360096,
|
|
"num_tokens": 837163.0,
|
|
"step": 460
|
|
},
|
|
{
|
|
"entropy": 7.2431192874908445,
|
|
"epoch": 0.3852526926263463,
|
|
"grad_norm": 1.28125,
|
|
"learning_rate": 0.00023200000000000003,
|
|
"loss": 6.9504,
|
|
"mean_token_accuracy": 0.0881062887609005,
|
|
"num_tokens": 846261.0,
|
|
"step": 465
|
|
},
|
|
{
|
|
"entropy": 7.118913030624389,
|
|
"epoch": 0.38939519469759737,
|
|
"grad_norm": 1.234375,
|
|
"learning_rate": 0.00023449999999999998,
|
|
"loss": 6.8595,
|
|
"mean_token_accuracy": 0.08703678250312805,
|
|
"num_tokens": 855655.0,
|
|
"step": 470
|
|
},
|
|
{
|
|
"entropy": 7.264123439788818,
|
|
"epoch": 0.39353769676884837,
|
|
"grad_norm": 1.265625,
|
|
"learning_rate": 0.000237,
|
|
"loss": 7.011,
|
|
"mean_token_accuracy": 0.08637121096253395,
|
|
"num_tokens": 864980.0,
|
|
"step": 475
|
|
},
|
|
{
|
|
"entropy": 7.17798171043396,
|
|
"epoch": 0.3976801988400994,
|
|
"grad_norm": 1.0390625,
|
|
"learning_rate": 0.0002395,
|
|
"loss": 6.8449,
|
|
"mean_token_accuracy": 0.08888368755578994,
|
|
"num_tokens": 874625.0,
|
|
"step": 480
|
|
},
|
|
{
|
|
"entropy": 7.17276029586792,
|
|
"epoch": 0.40182270091135047,
|
|
"grad_norm": 1.265625,
|
|
"learning_rate": 0.000242,
|
|
"loss": 6.8597,
|
|
"mean_token_accuracy": 0.08416522145271302,
|
|
"num_tokens": 884104.0,
|
|
"step": 485
|
|
},
|
|
{
|
|
"entropy": 7.224177455902099,
|
|
"epoch": 0.40596520298260147,
|
|
"grad_norm": 1.1875,
|
|
"learning_rate": 0.0002445,
|
|
"loss": 6.9764,
|
|
"mean_token_accuracy": 0.08516203612089157,
|
|
"num_tokens": 893355.0,
|
|
"step": 490
|
|
},
|
|
{
|
|
"entropy": 7.136541414260864,
|
|
"epoch": 0.4101077050538525,
|
|
"grad_norm": 1.3125,
|
|
"learning_rate": 0.000247,
|
|
"loss": 6.8995,
|
|
"mean_token_accuracy": 0.08847377672791482,
|
|
"num_tokens": 901666.0,
|
|
"step": 495
|
|
},
|
|
{
|
|
"entropy": 7.1944825649261475,
|
|
"epoch": 0.4142502071251036,
|
|
"grad_norm": 1.3984375,
|
|
"learning_rate": 0.0002495,
|
|
"loss": 6.8392,
|
|
"mean_token_accuracy": 0.08776508644223213,
|
|
"num_tokens": 911038.0,
|
|
"step": 500
|
|
},
|
|
{
|
|
"epoch": 0.4142502071251036,
|
|
"eval_entropy": 7.029010676117594,
|
|
"eval_loss": 6.951485633850098,
|
|
"eval_mean_token_accuracy": 0.08968717416625287,
|
|
"eval_num_tokens": 911038.0,
|
|
"eval_runtime": 2.9087,
|
|
"eval_samples_per_second": 1288.215,
|
|
"eval_steps_per_second": 161.242,
|
|
"step": 500
|
|
}
|
|
],
|
|
"logging_steps": 5,
|
|
"max_steps": 12060,
|
|
"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": 1346478667776000.0,
|
|
"train_batch_size": 16,
|
|
"trial_name": null,
|
|
"trial_params": null
|
|
}
|