2155 lines
64 KiB
JSON
2155 lines
64 KiB
JSON
|
{
|
||
|
|
"best_global_step": 300,
|
||
|
|
"best_metric": 0.17341697216033936,
|
||
|
|
"best_model_checkpoint": "/scratch/workspace/emma/standard-project-m2-mystery-machine/multilingual/results/checkpoints/sft_multilingual_v9_distilled/checkpoint-300",
|
||
|
|
"epoch": 1.6229438154240547,
|
||
|
|
"eval_steps": 100,
|
||
|
|
"global_step": 1900,
|
||
|
|
"is_hyper_param_search": false,
|
||
|
|
"is_local_process_zero": true,
|
||
|
|
"is_world_process_zero": true,
|
||
|
|
"log_history": [
|
||
|
|
{
|
||
|
|
"epoch": 0,
|
||
|
|
"eval_entropy": 0.0669629104218248,
|
||
|
|
"eval_loss": 4.374147415161133,
|
||
|
|
"eval_mean_token_accuracy": 0.684706018465761,
|
||
|
|
"eval_num_tokens": 0.0,
|
||
|
|
"eval_runtime": 129.5723,
|
||
|
|
"eval_samples_per_second": 25.515,
|
||
|
|
"eval_steps_per_second": 12.757,
|
||
|
|
"step": 0
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.24969844813458622,
|
||
|
|
"epoch": 0.008545182653279214,
|
||
|
|
"grad_norm": 6.274982929229736,
|
||
|
|
"learning_rate": 7.670454545454547e-07,
|
||
|
|
"loss": 1.2521350860595704,
|
||
|
|
"mean_token_accuracy": 0.764289478212595,
|
||
|
|
"num_tokens": 58678.0,
|
||
|
|
"step": 10
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.25262814382731447,
|
||
|
|
"epoch": 0.017090365306558428,
|
||
|
|
"grad_norm": 3.6314547061920166,
|
||
|
|
"learning_rate": 1.6193181818181818e-06,
|
||
|
|
"loss": 1.2871283531188964,
|
||
|
|
"mean_token_accuracy": 0.7669985197484493,
|
||
|
|
"num_tokens": 114530.0,
|
||
|
|
"step": 20
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.25956095702713355,
|
||
|
|
"epoch": 0.02563554795983764,
|
||
|
|
"grad_norm": 5.724814414978027,
|
||
|
|
"learning_rate": 2.4715909090909092e-06,
|
||
|
|
"loss": 1.3016060829162597,
|
||
|
|
"mean_token_accuracy": 0.7633212640881538,
|
||
|
|
"num_tokens": 168700.0,
|
||
|
|
"step": 30
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2609249378554523,
|
||
|
|
"epoch": 0.034180730613116855,
|
||
|
|
"grad_norm": 4.7607831954956055,
|
||
|
|
"learning_rate": 3.3238636363636366e-06,
|
||
|
|
"loss": 1.2199166297912598,
|
||
|
|
"mean_token_accuracy": 0.7618605919182301,
|
||
|
|
"num_tokens": 224044.0,
|
||
|
|
"step": 40
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2535087430849671,
|
||
|
|
"epoch": 0.04272591326639607,
|
||
|
|
"grad_norm": 2.8570611476898193,
|
||
|
|
"learning_rate": 4.176136363636364e-06,
|
||
|
|
"loss": 1.1118085861206055,
|
||
|
|
"mean_token_accuracy": 0.7751188822090626,
|
||
|
|
"num_tokens": 283946.0,
|
||
|
|
"step": 50
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.2823910743929446,
|
||
|
|
"epoch": 0.05127109591967528,
|
||
|
|
"grad_norm": 2.4557909965515137,
|
||
|
|
"learning_rate": 5.0284090909090905e-06,
|
||
|
|
"loss": 1.1430229187011718,
|
||
|
|
"mean_token_accuracy": 0.7865260414779186,
|
||
|
|
"num_tokens": 338744.0,
|
||
|
|
"step": 60
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.26092408254844485,
|
||
|
|
"epoch": 0.0598162785729545,
|
||
|
|
"grad_norm": 1.7713520526885986,
|
||
|
|
"learning_rate": 5.880681818181818e-06,
|
||
|
|
"loss": 1.0175163269042968,
|
||
|
|
"mean_token_accuracy": 0.8035369731485844,
|
||
|
|
"num_tokens": 392803.0,
|
||
|
|
"step": 70
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.28225490958648153,
|
||
|
|
"epoch": 0.06836146122623371,
|
||
|
|
"grad_norm": 1.8253531455993652,
|
||
|
|
"learning_rate": 6.732954545454545e-06,
|
||
|
|
"loss": 0.9419754028320313,
|
||
|
|
"mean_token_accuracy": 0.8099914431571961,
|
||
|
|
"num_tokens": 448031.0,
|
||
|
|
"step": 80
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.33381892020988746,
|
||
|
|
"epoch": 0.07690664387951293,
|
||
|
|
"grad_norm": 1.8320800065994263,
|
||
|
|
"learning_rate": 7.585227272727274e-06,
|
||
|
|
"loss": 0.9575636863708497,
|
||
|
|
"mean_token_accuracy": 0.803193487226963,
|
||
|
|
"num_tokens": 500928.0,
|
||
|
|
"step": 90
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.35758501158561556,
|
||
|
|
"epoch": 0.08545182653279214,
|
||
|
|
"grad_norm": 0.9949319958686829,
|
||
|
|
"learning_rate": 8.4375e-06,
|
||
|
|
"loss": 0.8450482368469239,
|
||
|
|
"mean_token_accuracy": 0.8077672526240349,
|
||
|
|
"num_tokens": 555919.0,
|
||
|
|
"step": 100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.08545182653279214,
|
||
|
|
"eval_entropy": 0.024625061223530184,
|
||
|
|
"eval_loss": 0.9070686101913452,
|
||
|
|
"eval_mean_token_accuracy": 0.8630697413134127,
|
||
|
|
"eval_num_tokens": 555919.0,
|
||
|
|
"eval_runtime": 130.7746,
|
||
|
|
"eval_samples_per_second": 25.28,
|
||
|
|
"eval_steps_per_second": 12.64,
|
||
|
|
"step": 100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.3740137527260231,
|
||
|
|
"epoch": 0.09399700918607136,
|
||
|
|
"grad_norm": 1.2319689989089966,
|
||
|
|
"learning_rate": 9.289772727272726e-06,
|
||
|
|
"loss": 0.643807315826416,
|
||
|
|
"mean_token_accuracy": 0.8233474880456925,
|
||
|
|
"num_tokens": 613943.0,
|
||
|
|
"step": 110
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4488209757953882,
|
||
|
|
"epoch": 0.10254219183935057,
|
||
|
|
"grad_norm": 0.8086826205253601,
|
||
|
|
"learning_rate": 1.0142045454545456e-05,
|
||
|
|
"loss": 0.688187837600708,
|
||
|
|
"mean_token_accuracy": 0.8218793146312237,
|
||
|
|
"num_tokens": 673784.0,
|
||
|
|
"step": 120
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.497951125074178,
|
||
|
|
"epoch": 0.11108737449262979,
|
||
|
|
"grad_norm": 0.9090407490730286,
|
||
|
|
"learning_rate": 1.0994318181818182e-05,
|
||
|
|
"loss": 0.6679890155792236,
|
||
|
|
"mean_token_accuracy": 0.8154753983020783,
|
||
|
|
"num_tokens": 727581.0,
|
||
|
|
"step": 130
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4890284799039364,
|
||
|
|
"epoch": 0.119632557145909,
|
||
|
|
"grad_norm": 0.6278298497200012,
|
||
|
|
"learning_rate": 1.184659090909091e-05,
|
||
|
|
"loss": 0.5986248016357422,
|
||
|
|
"mean_token_accuracy": 0.8232486955821514,
|
||
|
|
"num_tokens": 783574.0,
|
||
|
|
"step": 140
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5224538747221231,
|
||
|
|
"epoch": 0.1281777397991882,
|
||
|
|
"grad_norm": 0.5488609075546265,
|
||
|
|
"learning_rate": 1.2698863636363636e-05,
|
||
|
|
"loss": 0.6166930198669434,
|
||
|
|
"mean_token_accuracy": 0.8282795734703541,
|
||
|
|
"num_tokens": 843149.0,
|
||
|
|
"step": 150
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.563567110337317,
|
||
|
|
"epoch": 0.13672292245246742,
|
||
|
|
"grad_norm": 0.7587693333625793,
|
||
|
|
"learning_rate": 1.3551136363636364e-05,
|
||
|
|
"loss": 0.6144498348236084,
|
||
|
|
"mean_token_accuracy": 0.8195973120629787,
|
||
|
|
"num_tokens": 897263.0,
|
||
|
|
"step": 160
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5531872071325779,
|
||
|
|
"epoch": 0.14526810510574664,
|
||
|
|
"grad_norm": 0.4756774604320526,
|
||
|
|
"learning_rate": 1.4403409090909091e-05,
|
||
|
|
"loss": 0.5910183906555175,
|
||
|
|
"mean_token_accuracy": 0.8152173303067685,
|
||
|
|
"num_tokens": 956030.0,
|
||
|
|
"step": 170
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5073113800026476,
|
||
|
|
"epoch": 0.15381328775902586,
|
||
|
|
"grad_norm": 0.6585302352905273,
|
||
|
|
"learning_rate": 1.525568181818182e-05,
|
||
|
|
"loss": 0.5734929084777832,
|
||
|
|
"mean_token_accuracy": 0.8278453141450882,
|
||
|
|
"num_tokens": 1012316.0,
|
||
|
|
"step": 180
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5102395755238831,
|
||
|
|
"epoch": 0.16235847041230506,
|
||
|
|
"grad_norm": 0.5282315015792847,
|
||
|
|
"learning_rate": 1.6107954545454547e-05,
|
||
|
|
"loss": 0.5327517032623291,
|
||
|
|
"mean_token_accuracy": 0.836493419110775,
|
||
|
|
"num_tokens": 1072995.0,
|
||
|
|
"step": 190
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.522471790574491,
|
||
|
|
"epoch": 0.17090365306558428,
|
||
|
|
"grad_norm": 0.49718359112739563,
|
||
|
|
"learning_rate": 1.6960227272727273e-05,
|
||
|
|
"loss": 0.5635722160339356,
|
||
|
|
"mean_token_accuracy": 0.8290792062878609,
|
||
|
|
"num_tokens": 1128942.0,
|
||
|
|
"step": 200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.17090365306558428,
|
||
|
|
"eval_entropy": 0.14395250579487606,
|
||
|
|
"eval_loss": 0.22356745600700378,
|
||
|
|
"eval_mean_token_accuracy": 0.8651952759395999,
|
||
|
|
"eval_num_tokens": 1128942.0,
|
||
|
|
"eval_runtime": 130.1882,
|
||
|
|
"eval_samples_per_second": 25.394,
|
||
|
|
"eval_steps_per_second": 12.697,
|
||
|
|
"step": 200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5259293564595282,
|
||
|
|
"epoch": 0.1794488357188635,
|
||
|
|
"grad_norm": 0.6893606781959534,
|
||
|
|
"learning_rate": 1.78125e-05,
|
||
|
|
"loss": 0.5976882934570312,
|
||
|
|
"mean_token_accuracy": 0.8235495932400226,
|
||
|
|
"num_tokens": 1188900.0,
|
||
|
|
"step": 210
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5470972370356322,
|
||
|
|
"epoch": 0.18799401837214272,
|
||
|
|
"grad_norm": 0.4911622703075409,
|
||
|
|
"learning_rate": 1.866477272727273e-05,
|
||
|
|
"loss": 0.5525588512420654,
|
||
|
|
"mean_token_accuracy": 0.8309163197875022,
|
||
|
|
"num_tokens": 1249412.0,
|
||
|
|
"step": 220
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.496373859513551,
|
||
|
|
"epoch": 0.1965392010254219,
|
||
|
|
"grad_norm": 0.4767911732196808,
|
||
|
|
"learning_rate": 1.9517045454545455e-05,
|
||
|
|
"loss": 0.5422537326812744,
|
||
|
|
"mean_token_accuracy": 0.8364961229264736,
|
||
|
|
"num_tokens": 1304567.0,
|
||
|
|
"step": 230
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48757850332185626,
|
||
|
|
"epoch": 0.20508438367870113,
|
||
|
|
"grad_norm": 0.548505425453186,
|
||
|
|
"learning_rate": 2.036931818181818e-05,
|
||
|
|
"loss": 0.5599019527435303,
|
||
|
|
"mean_token_accuracy": 0.843613950163126,
|
||
|
|
"num_tokens": 1355876.0,
|
||
|
|
"step": 240
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5300518669188022,
|
||
|
|
"epoch": 0.21362956633198035,
|
||
|
|
"grad_norm": 0.6754305362701416,
|
||
|
|
"learning_rate": 2.122159090909091e-05,
|
||
|
|
"loss": 0.5412943840026856,
|
||
|
|
"mean_token_accuracy": 0.8387307159602642,
|
||
|
|
"num_tokens": 1407873.0,
|
||
|
|
"step": 250
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5064419998787344,
|
||
|
|
"epoch": 0.22217474898525957,
|
||
|
|
"grad_norm": 0.519078254699707,
|
||
|
|
"learning_rate": 2.2073863636363637e-05,
|
||
|
|
"loss": 0.5190999984741211,
|
||
|
|
"mean_token_accuracy": 0.8413542903959751,
|
||
|
|
"num_tokens": 1466071.0,
|
||
|
|
"step": 260
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.507786123547703,
|
||
|
|
"epoch": 0.23071993163853877,
|
||
|
|
"grad_norm": 0.6784747242927551,
|
||
|
|
"learning_rate": 2.2926136363636363e-05,
|
||
|
|
"loss": 0.5241509437561035,
|
||
|
|
"mean_token_accuracy": 0.8345548987388611,
|
||
|
|
"num_tokens": 1523657.0,
|
||
|
|
"step": 270
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4861731206998229,
|
||
|
|
"epoch": 0.239265114291818,
|
||
|
|
"grad_norm": 0.549146831035614,
|
||
|
|
"learning_rate": 2.3778409090909092e-05,
|
||
|
|
"loss": 0.4687436580657959,
|
||
|
|
"mean_token_accuracy": 0.8513894356787205,
|
||
|
|
"num_tokens": 1578449.0,
|
||
|
|
"step": 280
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4843604720197618,
|
||
|
|
"epoch": 0.2478102969450972,
|
||
|
|
"grad_norm": 0.45734071731567383,
|
||
|
|
"learning_rate": 2.463068181818182e-05,
|
||
|
|
"loss": 0.5005545616149902,
|
||
|
|
"mean_token_accuracy": 0.8470944315195084,
|
||
|
|
"num_tokens": 1642580.0,
|
||
|
|
"step": 290
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.49476482793688775,
|
||
|
|
"epoch": 0.2563554795983764,
|
||
|
|
"grad_norm": 0.42551544308662415,
|
||
|
|
"learning_rate": 2.5482954545454548e-05,
|
||
|
|
"loss": 0.5120632171630859,
|
||
|
|
"mean_token_accuracy": 0.8487209059298039,
|
||
|
|
"num_tokens": 1700114.0,
|
||
|
|
"step": 300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.2563554795983764,
|
||
|
|
"eval_entropy": 0.1536401255981387,
|
||
|
|
"eval_loss": 0.17341697216033936,
|
||
|
|
"eval_mean_token_accuracy": 0.9011337349289342,
|
||
|
|
"eval_num_tokens": 1700114.0,
|
||
|
|
"eval_runtime": 130.7274,
|
||
|
|
"eval_samples_per_second": 25.289,
|
||
|
|
"eval_steps_per_second": 12.645,
|
||
|
|
"step": 300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5316725081764162,
|
||
|
|
"epoch": 0.26490066225165565,
|
||
|
|
"grad_norm": 0.5366790890693665,
|
||
|
|
"learning_rate": 2.6335227272727274e-05,
|
||
|
|
"loss": 0.5439610958099366,
|
||
|
|
"mean_token_accuracy": 0.8321888409554958,
|
||
|
|
"num_tokens": 1756811.0,
|
||
|
|
"step": 310
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5158337057568133,
|
||
|
|
"epoch": 0.27344584490493484,
|
||
|
|
"grad_norm": 0.5545198321342468,
|
||
|
|
"learning_rate": 2.71875e-05,
|
||
|
|
"loss": 0.5202319145202636,
|
||
|
|
"mean_token_accuracy": 0.8420291461050511,
|
||
|
|
"num_tokens": 1814104.0,
|
||
|
|
"step": 320
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5089186687953771,
|
||
|
|
"epoch": 0.28199102755821404,
|
||
|
|
"grad_norm": 0.6484032869338989,
|
||
|
|
"learning_rate": 2.803977272727273e-05,
|
||
|
|
"loss": 0.5509964942932128,
|
||
|
|
"mean_token_accuracy": 0.8383300848305225,
|
||
|
|
"num_tokens": 1869014.0,
|
||
|
|
"step": 330
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5078026716597378,
|
||
|
|
"epoch": 0.2905362102114933,
|
||
|
|
"grad_norm": 0.5598916411399841,
|
||
|
|
"learning_rate": 2.8892045454545452e-05,
|
||
|
|
"loss": 0.5353553771972657,
|
||
|
|
"mean_token_accuracy": 0.8440746128559112,
|
||
|
|
"num_tokens": 1923809.0,
|
||
|
|
"step": 340
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5440114851109683,
|
||
|
|
"epoch": 0.2990813928647725,
|
||
|
|
"grad_norm": 0.5623669028282166,
|
||
|
|
"learning_rate": 2.9744318181818182e-05,
|
||
|
|
"loss": 0.5742122650146484,
|
||
|
|
"mean_token_accuracy": 0.8326552398502827,
|
||
|
|
"num_tokens": 1978798.0,
|
||
|
|
"step": 350
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5311623502522707,
|
||
|
|
"epoch": 0.3076265755180517,
|
||
|
|
"grad_norm": 0.5620706081390381,
|
||
|
|
"learning_rate": 2.9999637000233674e-05,
|
||
|
|
"loss": 0.5200393199920654,
|
||
|
|
"mean_token_accuracy": 0.8361483030021191,
|
||
|
|
"num_tokens": 2032103.0,
|
||
|
|
"step": 360
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5389031415805221,
|
||
|
|
"epoch": 0.3161717581713309,
|
||
|
|
"grad_norm": 0.7414209246635437,
|
||
|
|
"learning_rate": 2.9997859084489206e-05,
|
||
|
|
"loss": 0.5349567890167236,
|
||
|
|
"mean_token_accuracy": 0.8371621139347554,
|
||
|
|
"num_tokens": 2088855.0,
|
||
|
|
"step": 370
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.496375386044383,
|
||
|
|
"epoch": 0.3247169408246101,
|
||
|
|
"grad_norm": 0.5864424109458923,
|
||
|
|
"learning_rate": 2.999459975473429e-05,
|
||
|
|
"loss": 0.5306472778320312,
|
||
|
|
"mean_token_accuracy": 0.8493496462702751,
|
||
|
|
"num_tokens": 2145598.0,
|
||
|
|
"step": 380
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4972201234661043,
|
||
|
|
"epoch": 0.33326212347788936,
|
||
|
|
"grad_norm": 0.6628835797309875,
|
||
|
|
"learning_rate": 2.9989859332909326e-05,
|
||
|
|
"loss": 0.5208212852478027,
|
||
|
|
"mean_token_accuracy": 0.8440803237259388,
|
||
|
|
"num_tokens": 2201241.0,
|
||
|
|
"step": 390
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5351958076469601,
|
||
|
|
"epoch": 0.34180730613116855,
|
||
|
|
"grad_norm": 0.6935548186302185,
|
||
|
|
"learning_rate": 2.998363828724963e-05,
|
||
|
|
"loss": 0.5533876895904541,
|
||
|
|
"mean_token_accuracy": 0.8328421533107757,
|
||
|
|
"num_tokens": 2259084.0,
|
||
|
|
"step": 400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.34180730613116855,
|
||
|
|
"eval_entropy": 0.15067902416739043,
|
||
|
|
"eval_loss": 0.17932555079460144,
|
||
|
|
"eval_mean_token_accuracy": 0.9009451844514823,
|
||
|
|
"eval_num_tokens": 2259084.0,
|
||
|
|
"eval_runtime": 131.0499,
|
||
|
|
"eval_samples_per_second": 25.227,
|
||
|
|
"eval_steps_per_second": 12.614,
|
||
|
|
"step": 400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.529701460711658,
|
||
|
|
"epoch": 0.35035248878444775,
|
||
|
|
"grad_norm": 0.7940792441368103,
|
||
|
|
"learning_rate": 2.997593723223919e-05,
|
||
|
|
"loss": 0.5343834400177002,
|
||
|
|
"mean_token_accuracy": 0.8433877840638161,
|
||
|
|
"num_tokens": 2311600.0,
|
||
|
|
"step": 410
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5274290025234223,
|
||
|
|
"epoch": 0.358897671437727,
|
||
|
|
"grad_norm": 0.6897926330566406,
|
||
|
|
"learning_rate": 2.9966756928549957e-05,
|
||
|
|
"loss": 0.5001276969909668,
|
||
|
|
"mean_token_accuracy": 0.8402710266411304,
|
||
|
|
"num_tokens": 2374433.0,
|
||
|
|
"step": 420
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5436935693956911,
|
||
|
|
"epoch": 0.3674428540910062,
|
||
|
|
"grad_norm": 0.557988166809082,
|
||
|
|
"learning_rate": 2.995609828296672e-05,
|
||
|
|
"loss": 0.517430305480957,
|
||
|
|
"mean_token_accuracy": 0.8387674443423748,
|
||
|
|
"num_tokens": 2431718.0,
|
||
|
|
"step": 430
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.46528032384812834,
|
||
|
|
"epoch": 0.37598803674428544,
|
||
|
|
"grad_norm": 0.5683780908584595,
|
||
|
|
"learning_rate": 2.9943962348297538e-05,
|
||
|
|
"loss": 0.48638086318969725,
|
||
|
|
"mean_token_accuracy": 0.8531140975654126,
|
||
|
|
"num_tokens": 2487375.0,
|
||
|
|
"step": 440
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.45353667829185723,
|
||
|
|
"epoch": 0.38453321939756463,
|
||
|
|
"grad_norm": 1.0410420894622803,
|
||
|
|
"learning_rate": 2.993035032326975e-05,
|
||
|
|
"loss": 0.47857489585876467,
|
||
|
|
"mean_token_accuracy": 0.8567587099969387,
|
||
|
|
"num_tokens": 2542446.0,
|
||
|
|
"step": 450
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.500519336014986,
|
||
|
|
"epoch": 0.3930784020508438,
|
||
|
|
"grad_norm": 0.7571105360984802,
|
||
|
|
"learning_rate": 2.9915263552411558e-05,
|
||
|
|
"loss": 0.5148500919342041,
|
||
|
|
"mean_token_accuracy": 0.8473289743065834,
|
||
|
|
"num_tokens": 2599599.0,
|
||
|
|
"step": 460
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.44915042552165685,
|
||
|
|
"epoch": 0.40162358470412307,
|
||
|
|
"grad_norm": 0.43359076976776123,
|
||
|
|
"learning_rate": 2.989870352591924e-05,
|
||
|
|
"loss": 0.47877726554870603,
|
||
|
|
"mean_token_accuracy": 0.863776758313179,
|
||
|
|
"num_tokens": 2654964.0,
|
||
|
|
"step": 470
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4790582308545709,
|
||
|
|
"epoch": 0.41016876735740226,
|
||
|
|
"grad_norm": 0.7708096504211426,
|
||
|
|
"learning_rate": 2.988067187950994e-05,
|
||
|
|
"loss": 0.5002896308898925,
|
||
|
|
"mean_token_accuracy": 0.846701355278492,
|
||
|
|
"num_tokens": 2710272.0,
|
||
|
|
"step": 480
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4552692647092044,
|
||
|
|
"epoch": 0.41871395001068146,
|
||
|
|
"grad_norm": 0.5182470679283142,
|
||
|
|
"learning_rate": 2.9861170394260104e-05,
|
||
|
|
"loss": 0.45964322090148924,
|
||
|
|
"mean_token_accuracy": 0.8522793330252171,
|
||
|
|
"num_tokens": 2767882.0,
|
||
|
|
"step": 490
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48246400598436595,
|
||
|
|
"epoch": 0.4272591326639607,
|
||
|
|
"grad_norm": 0.5145803093910217,
|
||
|
|
"learning_rate": 2.9840200996429572e-05,
|
||
|
|
"loss": 0.49732532501220705,
|
||
|
|
"mean_token_accuracy": 0.8463977232575417,
|
||
|
|
"num_tokens": 2828122.0,
|
||
|
|
"step": 500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.4272591326639607,
|
||
|
|
"eval_entropy": 0.13816896783209395,
|
||
|
|
"eval_loss": 0.2536793053150177,
|
||
|
|
"eval_mean_token_accuracy": 0.8710268271813159,
|
||
|
|
"eval_num_tokens": 2828122.0,
|
||
|
|
"eval_runtime": 131.0714,
|
||
|
|
"eval_samples_per_second": 25.223,
|
||
|
|
"eval_steps_per_second": 12.611,
|
||
|
|
"step": 500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5290826369076967,
|
||
|
|
"epoch": 0.4358043153172399,
|
||
|
|
"grad_norm": 0.6639925241470337,
|
||
|
|
"learning_rate": 2.9817765757271272e-05,
|
||
|
|
"loss": 0.5043635368347168,
|
||
|
|
"mean_token_accuracy": 0.8379805460572243,
|
||
|
|
"num_tokens": 2886953.0,
|
||
|
|
"step": 510
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5226582545787096,
|
||
|
|
"epoch": 0.44434949797051915,
|
||
|
|
"grad_norm": 0.5005134344100952,
|
||
|
|
"learning_rate": 2.9793866892826682e-05,
|
||
|
|
"loss": 0.4872997760772705,
|
||
|
|
"mean_token_accuracy": 0.8422413349151612,
|
||
|
|
"num_tokens": 2947718.0,
|
||
|
|
"step": 520
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.532843924779445,
|
||
|
|
"epoch": 0.45289468062379834,
|
||
|
|
"grad_norm": 0.7342721223831177,
|
||
|
|
"learning_rate": 2.9768506763706892e-05,
|
||
|
|
"loss": 0.5164498805999755,
|
||
|
|
"mean_token_accuracy": 0.8450966902077198,
|
||
|
|
"num_tokens": 3005876.0,
|
||
|
|
"step": 530
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4689100462011993,
|
||
|
|
"epoch": 0.46143986327707753,
|
||
|
|
"grad_norm": 0.5711638927459717,
|
||
|
|
"learning_rate": 2.9741687874859468e-05,
|
||
|
|
"loss": 0.4971176624298096,
|
||
|
|
"mean_token_accuracy": 0.8549602694809437,
|
||
|
|
"num_tokens": 3058667.0,
|
||
|
|
"step": 540
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.49261180479079486,
|
||
|
|
"epoch": 0.4699850459303568,
|
||
|
|
"grad_norm": 0.5546977519989014,
|
||
|
|
"learning_rate": 2.9713412875321017e-05,
|
||
|
|
"loss": 0.4667799949645996,
|
||
|
|
"mean_token_accuracy": 0.8540357373654842,
|
||
|
|
"num_tokens": 3114766.0,
|
||
|
|
"step": 550
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47803974472917615,
|
||
|
|
"epoch": 0.478530228583636,
|
||
|
|
"grad_norm": 0.7118991017341614,
|
||
|
|
"learning_rate": 2.9683684557955518e-05,
|
||
|
|
"loss": 0.4712362766265869,
|
||
|
|
"mean_token_accuracy": 0.8569888420403003,
|
||
|
|
"num_tokens": 3169623.0,
|
||
|
|
"step": 560
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.46698532402515414,
|
||
|
|
"epoch": 0.48707541123691517,
|
||
|
|
"grad_norm": 0.7601621747016907,
|
||
|
|
"learning_rate": 2.9652505859178473e-05,
|
||
|
|
"loss": 0.48632082939147947,
|
||
|
|
"mean_token_accuracy": 0.8560906231403351,
|
||
|
|
"num_tokens": 3228520.0,
|
||
|
|
"step": 570
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5046033432707191,
|
||
|
|
"epoch": 0.4956205938901944,
|
||
|
|
"grad_norm": 0.705835223197937,
|
||
|
|
"learning_rate": 2.961987985866684e-05,
|
||
|
|
"loss": 0.5202618598937988,
|
||
|
|
"mean_token_accuracy": 0.8441015958786011,
|
||
|
|
"num_tokens": 3283437.0,
|
||
|
|
"step": 580
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4784255171660334,
|
||
|
|
"epoch": 0.5041657765434736,
|
||
|
|
"grad_norm": 0.9069309830665588,
|
||
|
|
"learning_rate": 2.958580977905488e-05,
|
||
|
|
"loss": 0.4908505916595459,
|
||
|
|
"mean_token_accuracy": 0.849098814278841,
|
||
|
|
"num_tokens": 3337297.0,
|
||
|
|
"step": 590
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5020056948065758,
|
||
|
|
"epoch": 0.5127109591967528,
|
||
|
|
"grad_norm": 0.6334726214408875,
|
||
|
|
"learning_rate": 2.955029898561578e-05,
|
||
|
|
"loss": 0.5153046607971191,
|
||
|
|
"mean_token_accuracy": 0.8445793382823468,
|
||
|
|
"num_tokens": 3395028.0,
|
||
|
|
"step": 600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.5127109591967528,
|
||
|
|
"eval_entropy": 0.1501043989929876,
|
||
|
|
"eval_loss": 0.18728521466255188,
|
||
|
|
"eval_mean_token_accuracy": 0.8962537333666737,
|
||
|
|
"eval_num_tokens": 3395028.0,
|
||
|
|
"eval_runtime": 130.641,
|
||
|
|
"eval_samples_per_second": 25.306,
|
||
|
|
"eval_steps_per_second": 12.653,
|
||
|
|
"step": 600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4977852301672101,
|
||
|
|
"epoch": 0.521256141850032,
|
||
|
|
"grad_norm": 0.7632060647010803,
|
||
|
|
"learning_rate": 2.9513350985929302e-05,
|
||
|
|
"loss": 0.47804670333862304,
|
||
|
|
"mean_token_accuracy": 0.8542221382260322,
|
||
|
|
"num_tokens": 3450430.0,
|
||
|
|
"step": 610
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4816632569767535,
|
||
|
|
"epoch": 0.5298013245033113,
|
||
|
|
"grad_norm": 0.7947196364402771,
|
||
|
|
"learning_rate": 2.94749694295353e-05,
|
||
|
|
"loss": 0.5373622417449951,
|
||
|
|
"mean_token_accuracy": 0.8464464962482452,
|
||
|
|
"num_tokens": 3505111.0,
|
||
|
|
"step": 620
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5585025523789227,
|
||
|
|
"epoch": 0.5383465071565905,
|
||
|
|
"grad_norm": 0.7158125638961792,
|
||
|
|
"learning_rate": 2.9435158107573226e-05,
|
||
|
|
"loss": 0.5493664741516113,
|
||
|
|
"mean_token_accuracy": 0.8305648066103458,
|
||
|
|
"num_tokens": 3559425.0,
|
||
|
|
"step": 630
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48834414817392824,
|
||
|
|
"epoch": 0.5468916898098697,
|
||
|
|
"grad_norm": 0.787824809551239,
|
||
|
|
"learning_rate": 2.939392095240769e-05,
|
||
|
|
"loss": 0.47605042457580565,
|
||
|
|
"mean_token_accuracy": 0.8558668762445449,
|
||
|
|
"num_tokens": 3617287.0,
|
||
|
|
"step": 640
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4972670005634427,
|
||
|
|
"epoch": 0.5554368724631489,
|
||
|
|
"grad_norm": 0.620593786239624,
|
||
|
|
"learning_rate": 2.9351262037239997e-05,
|
||
|
|
"loss": 0.5127807140350342,
|
||
|
|
"mean_token_accuracy": 0.8451636523008347,
|
||
|
|
"num_tokens": 3674109.0,
|
||
|
|
"step": 650
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.45892090993002055,
|
||
|
|
"epoch": 0.5639820551164281,
|
||
|
|
"grad_norm": 0.8967317938804626,
|
||
|
|
"learning_rate": 2.9307185575705873e-05,
|
||
|
|
"loss": 0.4700010299682617,
|
||
|
|
"mean_token_accuracy": 0.8574175283312797,
|
||
|
|
"num_tokens": 3731607.0,
|
||
|
|
"step": 660
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5267251801677049,
|
||
|
|
"epoch": 0.5725272377697074,
|
||
|
|
"grad_norm": 0.7435823678970337,
|
||
|
|
"learning_rate": 2.92616959214592e-05,
|
||
|
|
"loss": 0.5337582588195801,
|
||
|
|
"mean_token_accuracy": 0.8395570166409015,
|
||
|
|
"num_tokens": 3785709.0,
|
||
|
|
"step": 670
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48305057515390215,
|
||
|
|
"epoch": 0.5810724204229866,
|
||
|
|
"grad_norm": 0.6941148638725281,
|
||
|
|
"learning_rate": 2.9214797567742036e-05,
|
||
|
|
"loss": 0.4816096305847168,
|
||
|
|
"mean_token_accuracy": 0.8532864511013031,
|
||
|
|
"num_tokens": 3843573.0,
|
||
|
|
"step": 680
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4865381913725287,
|
||
|
|
"epoch": 0.5896176030762658,
|
||
|
|
"grad_norm": 0.6331520676612854,
|
||
|
|
"learning_rate": 2.9166495146940774e-05,
|
||
|
|
"loss": 0.49366283416748047,
|
||
|
|
"mean_token_accuracy": 0.8488726072013378,
|
||
|
|
"num_tokens": 3901058.0,
|
||
|
|
"step": 690
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5359061608090997,
|
||
|
|
"epoch": 0.598162785729545,
|
||
|
|
"grad_norm": 1.2058454751968384,
|
||
|
|
"learning_rate": 2.911679343012856e-05,
|
||
|
|
"loss": 0.5272064208984375,
|
||
|
|
"mean_token_accuracy": 0.8383475288748741,
|
||
|
|
"num_tokens": 3959864.0,
|
||
|
|
"step": 700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.598162785729545,
|
||
|
|
"eval_entropy": 0.13805914532756056,
|
||
|
|
"eval_loss": 0.2138196974992752,
|
||
|
|
"eval_mean_token_accuracy": 0.89399723147019,
|
||
|
|
"eval_num_tokens": 3959864.0,
|
||
|
|
"eval_runtime": 130.7947,
|
||
|
|
"eval_samples_per_second": 25.276,
|
||
|
|
"eval_steps_per_second": 12.638,
|
||
|
|
"step": 700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5001727496273816,
|
||
|
|
"epoch": 0.6067079683828241,
|
||
|
|
"grad_norm": 0.5492312908172607,
|
||
|
|
"learning_rate": 2.906569732659408e-05,
|
||
|
|
"loss": 0.5007120132446289,
|
||
|
|
"mean_token_accuracy": 0.8463241435587406,
|
||
|
|
"num_tokens": 4014798.0,
|
||
|
|
"step": 710
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.49296913351863625,
|
||
|
|
"epoch": 0.6152531510361035,
|
||
|
|
"grad_norm": 0.6929203271865845,
|
||
|
|
"learning_rate": 2.9013211883356568e-05,
|
||
|
|
"loss": 0.4912869453430176,
|
||
|
|
"mean_token_accuracy": 0.8495315484702587,
|
||
|
|
"num_tokens": 4069811.0,
|
||
|
|
"step": 720
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47721624858677386,
|
||
|
|
"epoch": 0.6237983336893826,
|
||
|
|
"grad_norm": 1.0193359851837158,
|
||
|
|
"learning_rate": 2.895934228466738e-05,
|
||
|
|
"loss": 0.5015659809112549,
|
||
|
|
"mean_token_accuracy": 0.848845724016428,
|
||
|
|
"num_tokens": 4124612.0,
|
||
|
|
"step": 730
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47406636411324143,
|
||
|
|
"epoch": 0.6323435163426618,
|
||
|
|
"grad_norm": 0.9448992013931274,
|
||
|
|
"learning_rate": 2.8904093851497842e-05,
|
||
|
|
"loss": 0.4977389335632324,
|
||
|
|
"mean_token_accuracy": 0.8561330914497376,
|
||
|
|
"num_tokens": 4178221.0,
|
||
|
|
"step": 740
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5023755330592394,
|
||
|
|
"epoch": 0.640888698995941,
|
||
|
|
"grad_norm": 0.6170384883880615,
|
||
|
|
"learning_rate": 2.88474720410137e-05,
|
||
|
|
"loss": 0.4966561794281006,
|
||
|
|
"mean_token_accuracy": 0.8498304672539234,
|
||
|
|
"num_tokens": 4238125.0,
|
||
|
|
"step": 750
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4658224150538445,
|
||
|
|
"epoch": 0.6494338816492202,
|
||
|
|
"grad_norm": 0.5623587965965271,
|
||
|
|
"learning_rate": 2.8789482446036095e-05,
|
||
|
|
"loss": 0.4847384452819824,
|
||
|
|
"mean_token_accuracy": 0.854324410110712,
|
||
|
|
"num_tokens": 4294105.0,
|
||
|
|
"step": 760
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5005891902372241,
|
||
|
|
"epoch": 0.6579790643024994,
|
||
|
|
"grad_norm": 0.6798438429832458,
|
||
|
|
"learning_rate": 2.8730130794489132e-05,
|
||
|
|
"loss": 0.49845376014709475,
|
||
|
|
"mean_token_accuracy": 0.8429157927632331,
|
||
|
|
"num_tokens": 4354422.0,
|
||
|
|
"step": 770
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5027361194137484,
|
||
|
|
"epoch": 0.6665242469557787,
|
||
|
|
"grad_norm": 0.8141709566116333,
|
||
|
|
"learning_rate": 2.8669422948834084e-05,
|
||
|
|
"loss": 0.508082914352417,
|
||
|
|
"mean_token_accuracy": 0.8500357121229172,
|
||
|
|
"num_tokens": 4408502.0,
|
||
|
|
"step": 780
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47243873588740826,
|
||
|
|
"epoch": 0.6750694296090579,
|
||
|
|
"grad_norm": 0.7849137783050537,
|
||
|
|
"learning_rate": 2.8607364905490345e-05,
|
||
|
|
"loss": 0.4844698429107666,
|
||
|
|
"mean_token_accuracy": 0.8605117335915565,
|
||
|
|
"num_tokens": 4464264.0,
|
||
|
|
"step": 790
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5196737521328032,
|
||
|
|
"epoch": 0.6836146122623371,
|
||
|
|
"grad_norm": 0.8575851917266846,
|
||
|
|
"learning_rate": 2.854396279424314e-05,
|
||
|
|
"loss": 0.5086193561553956,
|
||
|
|
"mean_token_accuracy": 0.846379928290844,
|
||
|
|
"num_tokens": 4523743.0,
|
||
|
|
"step": 800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.6836146122623371,
|
||
|
|
"eval_entropy": 0.12453013976739524,
|
||
|
|
"eval_loss": 0.2978106737136841,
|
||
|
|
"eval_mean_token_accuracy": 0.8744236879542028,
|
||
|
|
"eval_num_tokens": 4523743.0,
|
||
|
|
"eval_runtime": 130.3343,
|
||
|
|
"eval_samples_per_second": 25.366,
|
||
|
|
"eval_steps_per_second": 12.683,
|
||
|
|
"step": 800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5043912733905017,
|
||
|
|
"epoch": 0.6921597949156163,
|
||
|
|
"grad_norm": 0.8404580950737,
|
||
|
|
"learning_rate": 2.847922287763802e-05,
|
||
|
|
"loss": 0.5051683902740478,
|
||
|
|
"mean_token_accuracy": 0.8466556437313557,
|
||
|
|
"num_tokens": 4582699.0,
|
||
|
|
"step": 810
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5047137454617768,
|
||
|
|
"epoch": 0.7007049775688955,
|
||
|
|
"grad_norm": 0.7049486041069031,
|
||
|
|
"learning_rate": 2.841315155036232e-05,
|
||
|
|
"loss": 0.5086610317230225,
|
||
|
|
"mean_token_accuracy": 0.8485607713460922,
|
||
|
|
"num_tokens": 4637001.0,
|
||
|
|
"step": 820
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4949297777377069,
|
||
|
|
"epoch": 0.7092501602221748,
|
||
|
|
"grad_norm": 0.7774924635887146,
|
||
|
|
"learning_rate": 2.8345755338613512e-05,
|
||
|
|
"loss": 0.513433837890625,
|
||
|
|
"mean_token_accuracy": 0.8481399714946747,
|
||
|
|
"num_tokens": 4693463.0,
|
||
|
|
"step": 830
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4735300120897591,
|
||
|
|
"epoch": 0.717795342875454,
|
||
|
|
"grad_norm": 0.6502158045768738,
|
||
|
|
"learning_rate": 2.827704089945456e-05,
|
||
|
|
"loss": 0.47443389892578125,
|
||
|
|
"mean_token_accuracy": 0.8562471896409989,
|
||
|
|
"num_tokens": 4751136.0,
|
||
|
|
"step": 840
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48322221748530864,
|
||
|
|
"epoch": 0.7263405255287332,
|
||
|
|
"grad_norm": 0.6557545065879822,
|
||
|
|
"learning_rate": 2.820701502015639e-05,
|
||
|
|
"loss": 0.504110860824585,
|
||
|
|
"mean_token_accuracy": 0.8508345812559128,
|
||
|
|
"num_tokens": 4808609.0,
|
||
|
|
"step": 850
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5011111634783447,
|
||
|
|
"epoch": 0.7348857081820124,
|
||
|
|
"grad_norm": 0.5651575922966003,
|
||
|
|
"learning_rate": 2.813568461752746e-05,
|
||
|
|
"loss": 0.5229902744293213,
|
||
|
|
"mean_token_accuracy": 0.8521844178438187,
|
||
|
|
"num_tokens": 4866543.0,
|
||
|
|
"step": 860
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47421470619738104,
|
||
|
|
"epoch": 0.7434308908352916,
|
||
|
|
"grad_norm": 0.650168776512146,
|
||
|
|
"learning_rate": 2.8063056737230584e-05,
|
||
|
|
"loss": 0.48471846580505373,
|
||
|
|
"mean_token_accuracy": 0.8572772301733493,
|
||
|
|
"num_tokens": 4925117.0,
|
||
|
|
"step": 870
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47229467961005867,
|
||
|
|
"epoch": 0.7519760734885709,
|
||
|
|
"grad_norm": 0.6850579380989075,
|
||
|
|
"learning_rate": 2.7989138553086948e-05,
|
||
|
|
"loss": 0.5025987148284912,
|
||
|
|
"mean_token_accuracy": 0.8618568904697895,
|
||
|
|
"num_tokens": 4988704.0,
|
||
|
|
"step": 880
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5372770213522017,
|
||
|
|
"epoch": 0.7605212561418501,
|
||
|
|
"grad_norm": 0.5963494181632996,
|
||
|
|
"learning_rate": 2.7913937366367555e-05,
|
||
|
|
"loss": 0.5148426055908203,
|
||
|
|
"mean_token_accuracy": 0.8390157856047153,
|
||
|
|
"num_tokens": 5046909.0,
|
||
|
|
"step": 890
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5218867948744446,
|
||
|
|
"epoch": 0.7690664387951293,
|
||
|
|
"grad_norm": 0.8369334936141968,
|
||
|
|
"learning_rate": 2.7837460605072036e-05,
|
||
|
|
"loss": 0.5542621612548828,
|
||
|
|
"mean_token_accuracy": 0.8428195908665657,
|
||
|
|
"num_tokens": 5099629.0,
|
||
|
|
"step": 900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.7690664387951293,
|
||
|
|
"eval_entropy": 0.13862202610624463,
|
||
|
|
"eval_loss": 0.19889311492443085,
|
||
|
|
"eval_mean_token_accuracy": 0.8994441904303383,
|
||
|
|
"eval_num_tokens": 5099629.0,
|
||
|
|
"eval_runtime": 130.8862,
|
||
|
|
"eval_samples_per_second": 25.259,
|
||
|
|
"eval_steps_per_second": 12.629,
|
||
|
|
"step": 900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5125798557884991,
|
||
|
|
"epoch": 0.7776116214484085,
|
||
|
|
"grad_norm": 0.8306740522384644,
|
||
|
|
"learning_rate": 2.775971582319492e-05,
|
||
|
|
"loss": 0.5100288391113281,
|
||
|
|
"mean_token_accuracy": 0.8470876641571522,
|
||
|
|
"num_tokens": 5151115.0,
|
||
|
|
"step": 910
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4794777547009289,
|
||
|
|
"epoch": 0.7861568041016876,
|
||
|
|
"grad_norm": 0.6806477308273315,
|
||
|
|
"learning_rate": 2.768071069997953e-05,
|
||
|
|
"loss": 0.48182406425476076,
|
||
|
|
"mean_token_accuracy": 0.8533388070762158,
|
||
|
|
"num_tokens": 5210003.0,
|
||
|
|
"step": 920
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4697303189896047,
|
||
|
|
"epoch": 0.7947019867549668,
|
||
|
|
"grad_norm": 0.6518530249595642,
|
||
|
|
"learning_rate": 2.7600453039159433e-05,
|
||
|
|
"loss": 0.4823542594909668,
|
||
|
|
"mean_token_accuracy": 0.8549208372831345,
|
||
|
|
"num_tokens": 5267346.0,
|
||
|
|
"step": 930
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5235048852860927,
|
||
|
|
"epoch": 0.8032471694082461,
|
||
|
|
"grad_norm": 0.6777621507644653,
|
||
|
|
"learning_rate": 2.7518950768187643e-05,
|
||
|
|
"loss": 0.5249907970428467,
|
||
|
|
"mean_token_accuracy": 0.8450044021010399,
|
||
|
|
"num_tokens": 5320839.0,
|
||
|
|
"step": 940
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4732881594449282,
|
||
|
|
"epoch": 0.8117923520615253,
|
||
|
|
"grad_norm": 0.5677231550216675,
|
||
|
|
"learning_rate": 2.7436211937453582e-05,
|
||
|
|
"loss": 0.46576542854309083,
|
||
|
|
"mean_token_accuracy": 0.8635339729487896,
|
||
|
|
"num_tokens": 5377719.0,
|
||
|
|
"step": 950
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.471075529884547,
|
||
|
|
"epoch": 0.8203375347148045,
|
||
|
|
"grad_norm": 0.6974385976791382,
|
||
|
|
"learning_rate": 2.73522447194879e-05,
|
||
|
|
"loss": 0.46885366439819337,
|
||
|
|
"mean_token_accuracy": 0.8571250222623348,
|
||
|
|
"num_tokens": 5434122.0,
|
||
|
|
"step": 960
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4650014701299369,
|
||
|
|
"epoch": 0.8288827173680837,
|
||
|
|
"grad_norm": 0.7585994005203247,
|
||
|
|
"learning_rate": 2.726705740815523e-05,
|
||
|
|
"loss": 0.47807869911193845,
|
||
|
|
"mean_token_accuracy": 0.8536410674452781,
|
||
|
|
"num_tokens": 5485517.0,
|
||
|
|
"step": 970
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4916405067779124,
|
||
|
|
"epoch": 0.8374279000213629,
|
||
|
|
"grad_norm": 0.6748117208480835,
|
||
|
|
"learning_rate": 2.7180658417834977e-05,
|
||
|
|
"loss": 0.490537166595459,
|
||
|
|
"mean_token_accuracy": 0.8473445050418377,
|
||
|
|
"num_tokens": 5543445.0,
|
||
|
|
"step": 980
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4943195818923414,
|
||
|
|
"epoch": 0.8459730826746422,
|
||
|
|
"grad_norm": 1.2592843770980835,
|
||
|
|
"learning_rate": 2.7093056282590174e-05,
|
||
|
|
"loss": 0.48334059715270994,
|
||
|
|
"mean_token_accuracy": 0.8522989213466644,
|
||
|
|
"num_tokens": 5602103.0,
|
||
|
|
"step": 990
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5076028255745768,
|
||
|
|
"epoch": 0.8545182653279214,
|
||
|
|
"grad_norm": 1.219341516494751,
|
||
|
|
"learning_rate": 2.7004259655324522e-05,
|
||
|
|
"loss": 0.48604350090026854,
|
||
|
|
"mean_token_accuracy": 0.8460630945861339,
|
||
|
|
"num_tokens": 5661239.0,
|
||
|
|
"step": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.8545182653279214,
|
||
|
|
"eval_entropy": 0.12464745388029924,
|
||
|
|
"eval_loss": 0.3001876473426819,
|
||
|
|
"eval_mean_token_accuracy": 0.8766245750752061,
|
||
|
|
"eval_num_tokens": 5661239.0,
|
||
|
|
"eval_runtime": 130.0642,
|
||
|
|
"eval_samples_per_second": 25.418,
|
||
|
|
"eval_steps_per_second": 12.709,
|
||
|
|
"step": 1000
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4745888767298311,
|
||
|
|
"epoch": 0.8630634479812006,
|
||
|
|
"grad_norm": 0.7960503101348877,
|
||
|
|
"learning_rate": 2.6914277306927723e-05,
|
||
|
|
"loss": 0.4760587215423584,
|
||
|
|
"mean_token_accuracy": 0.8508720166981221,
|
||
|
|
"num_tokens": 5717472.0,
|
||
|
|
"step": 1010
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5108019274193794,
|
||
|
|
"epoch": 0.8716086306344798,
|
||
|
|
"grad_norm": 0.9308162927627563,
|
||
|
|
"learning_rate": 2.6823118125409113e-05,
|
||
|
|
"loss": 0.506138801574707,
|
||
|
|
"mean_token_accuracy": 0.8480221092700958,
|
||
|
|
"num_tokens": 5778978.0,
|
||
|
|
"step": 1020
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5153196372091771,
|
||
|
|
"epoch": 0.880153813287759,
|
||
|
|
"grad_norm": 0.6054751873016357,
|
||
|
|
"learning_rate": 2.6730791115019766e-05,
|
||
|
|
"loss": 0.4986146926879883,
|
||
|
|
"mean_token_accuracy": 0.843864256888628,
|
||
|
|
"num_tokens": 5838124.0,
|
||
|
|
"step": 1030
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.42288647515233607,
|
||
|
|
"epoch": 0.8886989959410383,
|
||
|
|
"grad_norm": 0.7642915844917297,
|
||
|
|
"learning_rate": 2.6637305395363077e-05,
|
||
|
|
"loss": 0.4333168029785156,
|
||
|
|
"mean_token_accuracy": 0.8680676862597465,
|
||
|
|
"num_tokens": 5894865.0,
|
||
|
|
"step": 1040
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47125883065164087,
|
||
|
|
"epoch": 0.8972441785943175,
|
||
|
|
"grad_norm": 0.8135762810707092,
|
||
|
|
"learning_rate": 2.6542670200493997e-05,
|
||
|
|
"loss": 0.48235602378845216,
|
||
|
|
"mean_token_accuracy": 0.855044524371624,
|
||
|
|
"num_tokens": 5949888.0,
|
||
|
|
"step": 1050
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4908870419487357,
|
||
|
|
"epoch": 0.9057893612475967,
|
||
|
|
"grad_norm": 0.7335505485534668,
|
||
|
|
"learning_rate": 2.6446894878006922e-05,
|
||
|
|
"loss": 0.5017701625823975,
|
||
|
|
"mean_token_accuracy": 0.8503725364804268,
|
||
|
|
"num_tokens": 6005188.0,
|
||
|
|
"step": 1060
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4389715285971761,
|
||
|
|
"epoch": 0.9143345439008759,
|
||
|
|
"grad_norm": 0.6651920080184937,
|
||
|
|
"learning_rate": 2.6349988888112396e-05,
|
||
|
|
"loss": 0.43928871154785154,
|
||
|
|
"mean_token_accuracy": 0.8601957596838474,
|
||
|
|
"num_tokens": 6060312.0,
|
||
|
|
"step": 1070
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4862962792627513,
|
||
|
|
"epoch": 0.9228797265541551,
|
||
|
|
"grad_norm": 0.8984395265579224,
|
||
|
|
"learning_rate": 2.6251961802702663e-05,
|
||
|
|
"loss": 0.47944092750549316,
|
||
|
|
"mean_token_accuracy": 0.8532074213027954,
|
||
|
|
"num_tokens": 6115090.0,
|
||
|
|
"step": 1080
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4814916422590613,
|
||
|
|
"epoch": 0.9314249092074343,
|
||
|
|
"grad_norm": 0.8483989834785461,
|
||
|
|
"learning_rate": 2.615282330440622e-05,
|
||
|
|
"loss": 0.5311069965362549,
|
||
|
|
"mean_token_accuracy": 0.8532755330204964,
|
||
|
|
"num_tokens": 6168388.0,
|
||
|
|
"step": 1090
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4910143165849149,
|
||
|
|
"epoch": 0.9399700918607136,
|
||
|
|
"grad_norm": 0.7318411469459534,
|
||
|
|
"learning_rate": 2.605258318563141e-05,
|
||
|
|
"loss": 0.4997721672058105,
|
||
|
|
"mean_token_accuracy": 0.8490157388150692,
|
||
|
|
"num_tokens": 6224395.0,
|
||
|
|
"step": 1100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 0.9399700918607136,
|
||
|
|
"eval_entropy": 0.13835578229947662,
|
||
|
|
"eval_loss": 0.2274738848209381,
|
||
|
|
"eval_mean_token_accuracy": 0.8866473511717785,
|
||
|
|
"eval_num_tokens": 6224395.0,
|
||
|
|
"eval_runtime": 130.918,
|
||
|
|
"eval_samples_per_second": 25.252,
|
||
|
|
"eval_steps_per_second": 12.626,
|
||
|
|
"step": 1100
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.501779202092439,
|
||
|
|
"epoch": 0.9485152745139928,
|
||
|
|
"grad_norm": 0.7204264402389526,
|
||
|
|
"learning_rate": 2.595125134759916e-05,
|
||
|
|
"loss": 0.49029059410095216,
|
||
|
|
"mean_token_accuracy": 0.8483284503221512,
|
||
|
|
"num_tokens": 6289589.0,
|
||
|
|
"step": 1110
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4868917277082801,
|
||
|
|
"epoch": 0.957060457167272,
|
||
|
|
"grad_norm": 0.7394916415214539,
|
||
|
|
"learning_rate": 2.5848837799365016e-05,
|
||
|
|
"loss": 0.49480552673339845,
|
||
|
|
"mean_token_accuracy": 0.8514926642179489,
|
||
|
|
"num_tokens": 6348430.0,
|
||
|
|
"step": 1120
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4662583138328046,
|
||
|
|
"epoch": 0.9656056398205511,
|
||
|
|
"grad_norm": 0.8898572325706482,
|
||
|
|
"learning_rate": 2.5745352656830474e-05,
|
||
|
|
"loss": 0.5194206237792969,
|
||
|
|
"mean_token_accuracy": 0.8523955218493938,
|
||
|
|
"num_tokens": 6401360.0,
|
||
|
|
"step": 1130
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47484123511239884,
|
||
|
|
"epoch": 0.9741508224738303,
|
||
|
|
"grad_norm": 0.7611355185508728,
|
||
|
|
"learning_rate": 2.56408061417438e-05,
|
||
|
|
"loss": 0.4896934986114502,
|
||
|
|
"mean_token_accuracy": 0.8509967491030693,
|
||
|
|
"num_tokens": 6461663.0,
|
||
|
|
"step": 1140
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4686635933350772,
|
||
|
|
"epoch": 0.9826960051271096,
|
||
|
|
"grad_norm": 0.6035947799682617,
|
||
|
|
"learning_rate": 2.5535208580690367e-05,
|
||
|
|
"loss": 0.4885598659515381,
|
||
|
|
"mean_token_accuracy": 0.8593245312571526,
|
||
|
|
"num_tokens": 6519420.0,
|
||
|
|
"step": 1150
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48161451891064644,
|
||
|
|
"epoch": 0.9912411877803888,
|
||
|
|
"grad_norm": 0.792752742767334,
|
||
|
|
"learning_rate": 2.542857040407263e-05,
|
||
|
|
"loss": 0.46279401779174806,
|
||
|
|
"mean_token_accuracy": 0.8534855127334595,
|
||
|
|
"num_tokens": 6578353.0,
|
||
|
|
"step": 1160
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4689274322707206,
|
||
|
|
"epoch": 0.999786370433668,
|
||
|
|
"grad_norm": 0.6302322745323181,
|
||
|
|
"learning_rate": 2.5320902145079904e-05,
|
||
|
|
"loss": 0.4713758945465088,
|
||
|
|
"mean_token_accuracy": 0.8566696748137475,
|
||
|
|
"num_tokens": 6635697.0,
|
||
|
|
"step": 1170
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4786196916892722,
|
||
|
|
"epoch": 1.0076906643879513,
|
||
|
|
"grad_norm": 0.51746666431427,
|
||
|
|
"learning_rate": 2.5212214438647915e-05,
|
||
|
|
"loss": 0.4642497539520264,
|
||
|
|
"mean_token_accuracy": 0.8600093875382397,
|
||
|
|
"num_tokens": 6688871.0,
|
||
|
|
"step": 1180
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4581842006184161,
|
||
|
|
"epoch": 1.0162358470412305,
|
||
|
|
"grad_norm": 0.7004144787788391,
|
||
|
|
"learning_rate": 2.5102518020408322e-05,
|
||
|
|
"loss": 0.43164548873901365,
|
||
|
|
"mean_token_accuracy": 0.8613746367394924,
|
||
|
|
"num_tokens": 6750885.0,
|
||
|
|
"step": 1190
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.45251300698146224,
|
||
|
|
"epoch": 1.0247810296945097,
|
||
|
|
"grad_norm": 0.9701333045959473,
|
||
|
|
"learning_rate": 2.4991823725628365e-05,
|
||
|
|
"loss": 0.4641223430633545,
|
||
|
|
"mean_token_accuracy": 0.8617949962615967,
|
||
|
|
"num_tokens": 6808434.0,
|
||
|
|
"step": 1200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.0247810296945097,
|
||
|
|
"eval_entropy": 0.13692736449210266,
|
||
|
|
"eval_loss": 0.22801218926906586,
|
||
|
|
"eval_mean_token_accuracy": 0.8881127385392164,
|
||
|
|
"eval_num_tokens": 6808434.0,
|
||
|
|
"eval_runtime": 129.0843,
|
||
|
|
"eval_samples_per_second": 25.611,
|
||
|
|
"eval_steps_per_second": 12.806,
|
||
|
|
"step": 1200
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4405954882502556,
|
||
|
|
"epoch": 1.0333262123477889,
|
||
|
|
"grad_norm": 0.5956438183784485,
|
||
|
|
"learning_rate": 2.4880142488140538e-05,
|
||
|
|
"loss": 0.43787102699279784,
|
||
|
|
"mean_token_accuracy": 0.8645174726843834,
|
||
|
|
"num_tokens": 6870293.0,
|
||
|
|
"step": 1210
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4427837684750557,
|
||
|
|
"epoch": 1.041871395001068,
|
||
|
|
"grad_norm": 0.7903404235839844,
|
||
|
|
"learning_rate": 2.4767485339262644e-05,
|
||
|
|
"loss": 0.44173803329467776,
|
||
|
|
"mean_token_accuracy": 0.8603828452527523,
|
||
|
|
"num_tokens": 6926016.0,
|
||
|
|
"step": 1220
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4276289961766452,
|
||
|
|
"epoch": 1.0504165776543473,
|
||
|
|
"grad_norm": 0.7248557806015015,
|
||
|
|
"learning_rate": 2.465386340670816e-05,
|
||
|
|
"loss": 0.4199423313140869,
|
||
|
|
"mean_token_accuracy": 0.8692133329808712,
|
||
|
|
"num_tokens": 6984273.0,
|
||
|
|
"step": 1230
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.501792199513875,
|
||
|
|
"epoch": 1.0589617603076267,
|
||
|
|
"grad_norm": 0.825380802154541,
|
||
|
|
"learning_rate": 2.4539287913487115e-05,
|
||
|
|
"loss": 0.5076761245727539,
|
||
|
|
"mean_token_accuracy": 0.8480165906250476,
|
||
|
|
"num_tokens": 7036425.0,
|
||
|
|
"step": 1240
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4217686806805432,
|
||
|
|
"epoch": 1.0675069429609059,
|
||
|
|
"grad_norm": 0.5441094636917114,
|
||
|
|
"learning_rate": 2.4423770176797514e-05,
|
||
|
|
"loss": 0.4357625961303711,
|
||
|
|
"mean_token_accuracy": 0.8657234266400338,
|
||
|
|
"num_tokens": 7093265.0,
|
||
|
|
"step": 1250
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.49576029628515245,
|
||
|
|
"epoch": 1.076052125614185,
|
||
|
|
"grad_norm": 0.5881739258766174,
|
||
|
|
"learning_rate": 2.430732160690748e-05,
|
||
|
|
"loss": 0.48982744216918944,
|
||
|
|
"mean_token_accuracy": 0.8510258793830872,
|
||
|
|
"num_tokens": 7148149.0,
|
||
|
|
"step": 1260
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4816284070722759,
|
||
|
|
"epoch": 1.0845973082674643,
|
||
|
|
"grad_norm": 1.089482307434082,
|
||
|
|
"learning_rate": 2.418995370602823e-05,
|
||
|
|
"loss": 0.4853186130523682,
|
||
|
|
"mean_token_accuracy": 0.8528147667646409,
|
||
|
|
"num_tokens": 7202320.0,
|
||
|
|
"step": 1270
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48163664937019346,
|
||
|
|
"epoch": 1.0931424909207434,
|
||
|
|
"grad_norm": 0.6645478010177612,
|
||
|
|
"learning_rate": 2.4071678067177934e-05,
|
||
|
|
"loss": 0.47566900253295896,
|
||
|
|
"mean_token_accuracy": 0.8495672412216664,
|
||
|
|
"num_tokens": 7260536.0,
|
||
|
|
"step": 1280
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4494845014065504,
|
||
|
|
"epoch": 1.1016876735740226,
|
||
|
|
"grad_norm": 0.8175600171089172,
|
||
|
|
"learning_rate": 2.395250637303659e-05,
|
||
|
|
"loss": 0.45380330085754395,
|
||
|
|
"mean_token_accuracy": 0.8638614401221275,
|
||
|
|
"num_tokens": 7322500.0,
|
||
|
|
"step": 1290
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.45753056965768335,
|
||
|
|
"epoch": 1.1102328562273018,
|
||
|
|
"grad_norm": 0.9088425636291504,
|
||
|
|
"learning_rate": 2.38324503947921e-05,
|
||
|
|
"loss": 0.4864628314971924,
|
||
|
|
"mean_token_accuracy": 0.855690510571003,
|
||
|
|
"num_tokens": 7378938.0,
|
||
|
|
"step": 1300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.1102328562273018,
|
||
|
|
"eval_entropy": 0.12984113263375313,
|
||
|
|
"eval_loss": 0.23559562861919403,
|
||
|
|
"eval_mean_token_accuracy": 0.891427369097544,
|
||
|
|
"eval_num_tokens": 7378938.0,
|
||
|
|
"eval_runtime": 130.8488,
|
||
|
|
"eval_samples_per_second": 25.266,
|
||
|
|
"eval_steps_per_second": 12.633,
|
||
|
|
"step": 1300
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.504523930279538,
|
||
|
|
"epoch": 1.118778038880581,
|
||
|
|
"grad_norm": 0.9652566909790039,
|
||
|
|
"learning_rate": 2.3711521990977556e-05,
|
||
|
|
"loss": 0.4652529716491699,
|
||
|
|
"mean_token_accuracy": 0.8536149144172669,
|
||
|
|
"num_tokens": 7437551.0,
|
||
|
|
"step": 1310
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4322365306201391,
|
||
|
|
"epoch": 1.1273232215338602,
|
||
|
|
"grad_norm": 0.982245922088623,
|
||
|
|
"learning_rate": 2.35897331062999e-05,
|
||
|
|
"loss": 0.45760526657104494,
|
||
|
|
"mean_token_accuracy": 0.8677045114338398,
|
||
|
|
"num_tokens": 7494605.0,
|
||
|
|
"step": 1320
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4661253343336284,
|
||
|
|
"epoch": 1.1358684041871394,
|
||
|
|
"grad_norm": 0.7378689646720886,
|
||
|
|
"learning_rate": 2.3467095770460116e-05,
|
||
|
|
"loss": 0.4972656726837158,
|
||
|
|
"mean_token_accuracy": 0.858714984357357,
|
||
|
|
"num_tokens": 7550459.0,
|
||
|
|
"step": 1330
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5022912069223822,
|
||
|
|
"epoch": 1.1444135868404186,
|
||
|
|
"grad_norm": 0.7334638833999634,
|
||
|
|
"learning_rate": 2.334362209696497e-05,
|
||
|
|
"loss": 0.4819185733795166,
|
||
|
|
"mean_token_accuracy": 0.8520683631300926,
|
||
|
|
"num_tokens": 7602163.0,
|
||
|
|
"step": 1340
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5100127911195159,
|
||
|
|
"epoch": 1.152958769493698,
|
||
|
|
"grad_norm": 0.7233748435974121,
|
||
|
|
"learning_rate": 2.321932428193051e-05,
|
||
|
|
"loss": 0.4884189605712891,
|
||
|
|
"mean_token_accuracy": 0.85202471986413,
|
||
|
|
"num_tokens": 7661593.0,
|
||
|
|
"step": 1350
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4745957013219595,
|
||
|
|
"epoch": 1.1615039521469772,
|
||
|
|
"grad_norm": 0.6802652478218079,
|
||
|
|
"learning_rate": 2.3094214602877382e-05,
|
||
|
|
"loss": 0.45793399810791013,
|
||
|
|
"mean_token_accuracy": 0.8596083417534828,
|
||
|
|
"num_tokens": 7718576.0,
|
||
|
|
"step": 1360
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47644422790035607,
|
||
|
|
"epoch": 1.1700491348002564,
|
||
|
|
"grad_norm": 0.9927780628204346,
|
||
|
|
"learning_rate": 2.2968305417518144e-05,
|
||
|
|
"loss": 0.4342164993286133,
|
||
|
|
"mean_token_accuracy": 0.8588686197996139,
|
||
|
|
"num_tokens": 7774393.0,
|
||
|
|
"step": 1370
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.46737509220838547,
|
||
|
|
"epoch": 1.1785943174535356,
|
||
|
|
"grad_norm": 0.9275594353675842,
|
||
|
|
"learning_rate": 2.2841609162536596e-05,
|
||
|
|
"loss": 0.4524184226989746,
|
||
|
|
"mean_token_accuracy": 0.8595091588795185,
|
||
|
|
"num_tokens": 7833069.0,
|
||
|
|
"step": 1380
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.466662473231554,
|
||
|
|
"epoch": 1.1871395001068148,
|
||
|
|
"grad_norm": 0.6475253701210022,
|
||
|
|
"learning_rate": 2.271413835235938e-05,
|
||
|
|
"loss": 0.4612546443939209,
|
||
|
|
"mean_token_accuracy": 0.8563490323722363,
|
||
|
|
"num_tokens": 7892947.0,
|
||
|
|
"step": 1390
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48723165187984707,
|
||
|
|
"epoch": 1.195684682760094,
|
||
|
|
"grad_norm": 0.8730847239494324,
|
||
|
|
"learning_rate": 2.2585905577919837e-05,
|
||
|
|
"loss": 0.45258269309997556,
|
||
|
|
"mean_token_accuracy": 0.8530362658202648,
|
||
|
|
"num_tokens": 7951122.0,
|
||
|
|
"step": 1400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.195684682760094,
|
||
|
|
"eval_entropy": 0.13191182531880669,
|
||
|
|
"eval_loss": 0.25020337104797363,
|
||
|
|
"eval_mean_token_accuracy": 0.8841259847319061,
|
||
|
|
"eval_num_tokens": 7951122.0,
|
||
|
|
"eval_runtime": 130.7361,
|
||
|
|
"eval_samples_per_second": 25.288,
|
||
|
|
"eval_steps_per_second": 12.644,
|
||
|
|
"step": 1400
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4679116272833198,
|
||
|
|
"epoch": 1.2042298654133732,
|
||
|
|
"grad_norm": 0.9716050624847412,
|
||
|
|
"learning_rate": 2.2456923505414366e-05,
|
||
|
|
"loss": 0.45508151054382323,
|
||
|
|
"mean_token_accuracy": 0.8529311671853066,
|
||
|
|
"num_tokens": 8006193.0,
|
||
|
|
"step": 1410
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4207824946381152,
|
||
|
|
"epoch": 1.2127750480666524,
|
||
|
|
"grad_norm": 0.9770030379295349,
|
||
|
|
"learning_rate": 2.23272048750513e-05,
|
||
|
|
"loss": 0.4458351135253906,
|
||
|
|
"mean_token_accuracy": 0.8706255331635475,
|
||
|
|
"num_tokens": 8058758.0,
|
||
|
|
"step": 1420
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.41676688510924575,
|
||
|
|
"epoch": 1.2213202307199316,
|
||
|
|
"grad_norm": 0.693558931350708,
|
||
|
|
"learning_rate": 2.2196762499792502e-05,
|
||
|
|
"loss": 0.45579938888549804,
|
||
|
|
"mean_token_accuracy": 0.8652678288519382,
|
||
|
|
"num_tokens": 8115947.0,
|
||
|
|
"step": 1430
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.49633874078281226,
|
||
|
|
"epoch": 1.2298654133732108,
|
||
|
|
"grad_norm": 0.7726420164108276,
|
||
|
|
"learning_rate": 2.2065609264087755e-05,
|
||
|
|
"loss": 0.4634819984436035,
|
||
|
|
"mean_token_accuracy": 0.8566023781895638,
|
||
|
|
"num_tokens": 8174159.0,
|
||
|
|
"step": 1440
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4586691297125071,
|
||
|
|
"epoch": 1.23841059602649,
|
||
|
|
"grad_norm": 0.9443950057029724,
|
||
|
|
"learning_rate": 2.1933758122602112e-05,
|
||
|
|
"loss": 0.44040789604187014,
|
||
|
|
"mean_token_accuracy": 0.8588434293866157,
|
||
|
|
"num_tokens": 8230317.0,
|
||
|
|
"step": 1450
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4659241737797856,
|
||
|
|
"epoch": 1.2469557786797694,
|
||
|
|
"grad_norm": 0.6670970320701599,
|
||
|
|
"learning_rate": 2.180122209893628e-05,
|
||
|
|
"loss": 0.4633673667907715,
|
||
|
|
"mean_token_accuracy": 0.8570841707289218,
|
||
|
|
"num_tokens": 8286279.0,
|
||
|
|
"step": 1460
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.437501065293327,
|
||
|
|
"epoch": 1.2555009613330486,
|
||
|
|
"grad_norm": 0.8437661528587341,
|
||
|
|
"learning_rate": 2.1668014284340244e-05,
|
||
|
|
"loss": 0.43558216094970703,
|
||
|
|
"mean_token_accuracy": 0.8636670149862766,
|
||
|
|
"num_tokens": 8344641.0,
|
||
|
|
"step": 1470
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4582040155306458,
|
||
|
|
"epoch": 1.2640461439863278,
|
||
|
|
"grad_norm": 0.7960573434829712,
|
||
|
|
"learning_rate": 2.1534147836420132e-05,
|
||
|
|
"loss": 0.46987128257751465,
|
||
|
|
"mean_token_accuracy": 0.8601755306124688,
|
||
|
|
"num_tokens": 8399645.0,
|
||
|
|
"step": 1480
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4295470048673451,
|
||
|
|
"epoch": 1.272591326639607,
|
||
|
|
"grad_norm": 0.8968801498413086,
|
||
|
|
"learning_rate": 2.1399635977838633e-05,
|
||
|
|
"loss": 0.4409494876861572,
|
||
|
|
"mean_token_accuracy": 0.8653147928416729,
|
||
|
|
"num_tokens": 8456116.0,
|
||
|
|
"step": 1490
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4707821879070252,
|
||
|
|
"epoch": 1.2811365092928861,
|
||
|
|
"grad_norm": 0.7065398693084717,
|
||
|
|
"learning_rate": 2.126449199500887e-05,
|
||
|
|
"loss": 0.4831379413604736,
|
||
|
|
"mean_token_accuracy": 0.8583334997296334,
|
||
|
|
"num_tokens": 8509846.0,
|
||
|
|
"step": 1500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.2811365092928861,
|
||
|
|
"eval_entropy": 0.13262949710406466,
|
||
|
|
"eval_loss": 0.20845149457454681,
|
||
|
|
"eval_mean_token_accuracy": 0.8964909433741607,
|
||
|
|
"eval_num_tokens": 8509846.0,
|
||
|
|
"eval_runtime": 130.3683,
|
||
|
|
"eval_samples_per_second": 25.359,
|
||
|
|
"eval_steps_per_second": 12.679,
|
||
|
|
"step": 1500
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4395508569199592,
|
||
|
|
"epoch": 1.2896816919461653,
|
||
|
|
"grad_norm": 0.7794111967086792,
|
||
|
|
"learning_rate": 2.1128729236782066e-05,
|
||
|
|
"loss": 0.4607836246490479,
|
||
|
|
"mean_token_accuracy": 0.8647575780749321,
|
||
|
|
"num_tokens": 8565953.0,
|
||
|
|
"step": 1510
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4764142186846584,
|
||
|
|
"epoch": 1.2982268745994445,
|
||
|
|
"grad_norm": 1.164459466934204,
|
||
|
|
"learning_rate": 2.0992361113129005e-05,
|
||
|
|
"loss": 0.4883578777313232,
|
||
|
|
"mean_token_accuracy": 0.8497100658714771,
|
||
|
|
"num_tokens": 8625215.0,
|
||
|
|
"step": 1520
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.492379231704399,
|
||
|
|
"epoch": 1.3067720572527237,
|
||
|
|
"grad_norm": 0.9724196195602417,
|
||
|
|
"learning_rate": 2.0855401093815458e-05,
|
||
|
|
"loss": 0.4727339267730713,
|
||
|
|
"mean_token_accuracy": 0.8533373884856701,
|
||
|
|
"num_tokens": 8677802.0,
|
||
|
|
"step": 1530
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4452589777763933,
|
||
|
|
"epoch": 1.315317239906003,
|
||
|
|
"grad_norm": 0.9151174426078796,
|
||
|
|
"learning_rate": 2.0717862707071716e-05,
|
||
|
|
"loss": 0.4631012439727783,
|
||
|
|
"mean_token_accuracy": 0.8583762601017952,
|
||
|
|
"num_tokens": 8732079.0,
|
||
|
|
"step": 1540
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4605603158473969,
|
||
|
|
"epoch": 1.3238624225592823,
|
||
|
|
"grad_norm": 0.7729523777961731,
|
||
|
|
"learning_rate": 2.0579759538256323e-05,
|
||
|
|
"loss": 0.4707770347595215,
|
||
|
|
"mean_token_accuracy": 0.8546851448714733,
|
||
|
|
"num_tokens": 8786563.0,
|
||
|
|
"step": 1550
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4637235806323588,
|
||
|
|
"epoch": 1.3324076052125613,
|
||
|
|
"grad_norm": 0.9665090441703796,
|
||
|
|
"learning_rate": 2.044110522851419e-05,
|
||
|
|
"loss": 0.45175724029541015,
|
||
|
|
"mean_token_accuracy": 0.8577323235571385,
|
||
|
|
"num_tokens": 8850096.0,
|
||
|
|
"step": 1560
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.43333199685439466,
|
||
|
|
"epoch": 1.3409527878658407,
|
||
|
|
"grad_norm": 1.0164531469345093,
|
||
|
|
"learning_rate": 2.0301913473429193e-05,
|
||
|
|
"loss": 0.45435385704040526,
|
||
|
|
"mean_token_accuracy": 0.8607866078615188,
|
||
|
|
"num_tokens": 8906790.0,
|
||
|
|
"step": 1570
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47894956804811956,
|
||
|
|
"epoch": 1.34949797051912,
|
||
|
|
"grad_norm": 0.9175162315368652,
|
||
|
|
"learning_rate": 2.01621980216714e-05,
|
||
|
|
"loss": 0.4707216262817383,
|
||
|
|
"mean_token_accuracy": 0.8571217954158783,
|
||
|
|
"num_tokens": 8962103.0,
|
||
|
|
"step": 1580
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4955364347435534,
|
||
|
|
"epoch": 1.358043153172399,
|
||
|
|
"grad_norm": 0.7004621624946594,
|
||
|
|
"learning_rate": 2.002197267363901e-05,
|
||
|
|
"loss": 0.4925191402435303,
|
||
|
|
"mean_token_accuracy": 0.8490186050534249,
|
||
|
|
"num_tokens": 9015670.0,
|
||
|
|
"step": 1590
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.44045674256049094,
|
||
|
|
"epoch": 1.3665883358256783,
|
||
|
|
"grad_norm": 0.78447425365448,
|
||
|
|
"learning_rate": 1.9881251280095264e-05,
|
||
|
|
"loss": 0.44242281913757325,
|
||
|
|
"mean_token_accuracy": 0.8647661522030831,
|
||
|
|
"num_tokens": 9070057.0,
|
||
|
|
"step": 1600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.3665883358256783,
|
||
|
|
"eval_entropy": 0.13024739608736524,
|
||
|
|
"eval_loss": 0.24764536321163177,
|
||
|
|
"eval_mean_token_accuracy": 0.8851719131919736,
|
||
|
|
"eval_num_tokens": 9070057.0,
|
||
|
|
"eval_runtime": 131.4618,
|
||
|
|
"eval_samples_per_second": 25.148,
|
||
|
|
"eval_steps_per_second": 12.574,
|
||
|
|
"step": 1600
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4327243532519788,
|
||
|
|
"epoch": 1.3751335184789575,
|
||
|
|
"grad_norm": 0.8228652477264404,
|
||
|
|
"learning_rate": 1.974004774080031e-05,
|
||
|
|
"loss": 0.4115453243255615,
|
||
|
|
"mean_token_accuracy": 0.8628325082361699,
|
||
|
|
"num_tokens": 9128844.0,
|
||
|
|
"step": 1610
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4597356241662055,
|
||
|
|
"epoch": 1.3836787011322367,
|
||
|
|
"grad_norm": 0.6152841448783875,
|
||
|
|
"learning_rate": 1.9598376003138247e-05,
|
||
|
|
"loss": 0.4410817623138428,
|
||
|
|
"mean_token_accuracy": 0.8625422030687332,
|
||
|
|
"num_tokens": 9186025.0,
|
||
|
|
"step": 1620
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48284322540275754,
|
||
|
|
"epoch": 1.3922238837855159,
|
||
|
|
"grad_norm": 0.9332865476608276,
|
||
|
|
"learning_rate": 1.9456250060739498e-05,
|
||
|
|
"loss": 0.4874548435211182,
|
||
|
|
"mean_token_accuracy": 0.8537392042577266,
|
||
|
|
"num_tokens": 9239454.0,
|
||
|
|
"step": 1630
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.45977171589620414,
|
||
|
|
"epoch": 1.400769066438795,
|
||
|
|
"grad_norm": 0.8458712100982666,
|
||
|
|
"learning_rate": 1.9313683952098566e-05,
|
||
|
|
"loss": 0.4500466823577881,
|
||
|
|
"mean_token_accuracy": 0.8578137546777725,
|
||
|
|
"num_tokens": 9294748.0,
|
||
|
|
"step": 1640
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.43004854638129475,
|
||
|
|
"epoch": 1.4093142490920743,
|
||
|
|
"grad_norm": 0.9201598763465881,
|
||
|
|
"learning_rate": 1.917069175918738e-05,
|
||
|
|
"loss": 0.4568291187286377,
|
||
|
|
"mean_token_accuracy": 0.8618893027305603,
|
||
|
|
"num_tokens": 9348879.0,
|
||
|
|
"step": 1650
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4350295281037688,
|
||
|
|
"epoch": 1.4178594317453537,
|
||
|
|
"grad_norm": 0.942304253578186,
|
||
|
|
"learning_rate": 1.902728760606438e-05,
|
||
|
|
"loss": 0.42952899932861327,
|
||
|
|
"mean_token_accuracy": 0.8623266451060771,
|
||
|
|
"num_tokens": 9406528.0,
|
||
|
|
"step": 1660
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.47831510389223697,
|
||
|
|
"epoch": 1.4264046143986326,
|
||
|
|
"grad_norm": 0.6833867430686951,
|
||
|
|
"learning_rate": 1.888348565747936e-05,
|
||
|
|
"loss": 0.4522396087646484,
|
||
|
|
"mean_token_accuracy": 0.857250015437603,
|
||
|
|
"num_tokens": 9469060.0,
|
||
|
|
"step": 1670
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4272033490706235,
|
||
|
|
"epoch": 1.434949797051912,
|
||
|
|
"grad_norm": 0.7562753558158875,
|
||
|
|
"learning_rate": 1.87393001174744e-05,
|
||
|
|
"loss": 0.4609020233154297,
|
||
|
|
"mean_token_accuracy": 0.8627648457884789,
|
||
|
|
"num_tokens": 9525896.0,
|
||
|
|
"step": 1680
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.43255489226430655,
|
||
|
|
"epoch": 1.4434949797051913,
|
||
|
|
"grad_norm": 0.9394037127494812,
|
||
|
|
"learning_rate": 1.8594745227980808e-05,
|
||
|
|
"loss": 0.42289090156555176,
|
||
|
|
"mean_token_accuracy": 0.8683697424829007,
|
||
|
|
"num_tokens": 9584033.0,
|
||
|
|
"step": 1690
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.46635143235325816,
|
||
|
|
"epoch": 1.4520401623584704,
|
||
|
|
"grad_norm": 0.9890229105949402,
|
||
|
|
"learning_rate": 1.8449835267412428e-05,
|
||
|
|
"loss": 0.4432528495788574,
|
||
|
|
"mean_token_accuracy": 0.8590702742338181,
|
||
|
|
"num_tokens": 9639121.0,
|
||
|
|
"step": 1700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.4520401623584704,
|
||
|
|
"eval_entropy": 0.13027989997540326,
|
||
|
|
"eval_loss": 0.21824395656585693,
|
||
|
|
"eval_mean_token_accuracy": 0.8969438583001323,
|
||
|
|
"eval_num_tokens": 9639121.0,
|
||
|
|
"eval_runtime": 131.2952,
|
||
|
|
"eval_samples_per_second": 25.18,
|
||
|
|
"eval_steps_per_second": 12.59,
|
||
|
|
"step": 1700
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4192576443310827,
|
||
|
|
"epoch": 1.4605853450117496,
|
||
|
|
"grad_norm": 0.7131622433662415,
|
||
|
|
"learning_rate": 1.8304584549255242e-05,
|
||
|
|
"loss": 0.42078628540039065,
|
||
|
|
"mean_token_accuracy": 0.8704721838235855,
|
||
|
|
"num_tokens": 9692205.0,
|
||
|
|
"step": 1710
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.41785974830854683,
|
||
|
|
"epoch": 1.4691305276650288,
|
||
|
|
"grad_norm": 0.6923640966415405,
|
||
|
|
"learning_rate": 1.815900742065358e-05,
|
||
|
|
"loss": 0.38985638618469237,
|
||
|
|
"mean_token_accuracy": 0.8742700718343258,
|
||
|
|
"num_tokens": 9754785.0,
|
||
|
|
"step": 1720
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4400124318897724,
|
||
|
|
"epoch": 1.477675710318308,
|
||
|
|
"grad_norm": 1.103582739830017,
|
||
|
|
"learning_rate": 1.801311826099297e-05,
|
||
|
|
"loss": 0.42641592025756836,
|
||
|
|
"mean_token_accuracy": 0.8634009219706058,
|
||
|
|
"num_tokens": 9811488.0,
|
||
|
|
"step": 1730
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.44537130240350964,
|
||
|
|
"epoch": 1.4862208929715872,
|
||
|
|
"grad_norm": 0.9648677110671997,
|
||
|
|
"learning_rate": 1.786693148047982e-05,
|
||
|
|
"loss": 0.4516180992126465,
|
||
|
|
"mean_token_accuracy": 0.8613857299089431,
|
||
|
|
"num_tokens": 9864884.0,
|
||
|
|
"step": 1740
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4132983233779669,
|
||
|
|
"epoch": 1.4947660756248664,
|
||
|
|
"grad_norm": 0.7255743741989136,
|
||
|
|
"learning_rate": 1.772046151871805e-05,
|
||
|
|
"loss": 0.4096498966217041,
|
||
|
|
"mean_token_accuracy": 0.8688085481524468,
|
||
|
|
"num_tokens": 9923274.0,
|
||
|
|
"step": 1750
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4117629977874458,
|
||
|
|
"epoch": 1.5033112582781456,
|
||
|
|
"grad_norm": 0.9515234231948853,
|
||
|
|
"learning_rate": 1.757372284328282e-05,
|
||
|
|
"loss": 0.43401236534118653,
|
||
|
|
"mean_token_accuracy": 0.8734201341867447,
|
||
|
|
"num_tokens": 9981945.0,
|
||
|
|
"step": 1760
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4250434666639194,
|
||
|
|
"epoch": 1.511856440931425,
|
||
|
|
"grad_norm": 0.8354512453079224,
|
||
|
|
"learning_rate": 1.7426729948291474e-05,
|
||
|
|
"loss": 0.43279552459716797,
|
||
|
|
"mean_token_accuracy": 0.8683175548911095,
|
||
|
|
"num_tokens": 10043784.0,
|
||
|
|
"step": 1770
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.42646104865707457,
|
||
|
|
"epoch": 1.520401623584704,
|
||
|
|
"grad_norm": 0.8140296936035156,
|
||
|
|
"learning_rate": 1.727949735297192e-05,
|
||
|
|
"loss": 0.4162501335144043,
|
||
|
|
"mean_token_accuracy": 0.8683593071997165,
|
||
|
|
"num_tokens": 10099326.0,
|
||
|
|
"step": 1780
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.44492706399178134,
|
||
|
|
"epoch": 1.5289468062379834,
|
||
|
|
"grad_norm": 1.0218313932418823,
|
||
|
|
"learning_rate": 1.713203960022847e-05,
|
||
|
|
"loss": 0.4697896480560303,
|
||
|
|
"mean_token_accuracy": 0.8601226754486561,
|
||
|
|
"num_tokens": 10155233.0,
|
||
|
|
"step": 1790
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.5192175327800215,
|
||
|
|
"epoch": 1.5374919888912626,
|
||
|
|
"grad_norm": 0.7575676441192627,
|
||
|
|
"learning_rate": 1.6984371255205355e-05,
|
||
|
|
"loss": 0.49686422348022463,
|
||
|
|
"mean_token_accuracy": 0.8497159123420716,
|
||
|
|
"num_tokens": 10214346.0,
|
||
|
|
"step": 1800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.5374919888912626,
|
||
|
|
"eval_entropy": 0.12461854687385392,
|
||
|
|
"eval_loss": 0.2748284637928009,
|
||
|
|
"eval_mean_token_accuracy": 0.8812604682126769,
|
||
|
|
"eval_num_tokens": 10214346.0,
|
||
|
|
"eval_runtime": 131.4892,
|
||
|
|
"eval_samples_per_second": 25.143,
|
||
|
|
"eval_steps_per_second": 12.571,
|
||
|
|
"step": 1800
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.45915504824370146,
|
||
|
|
"epoch": 1.5460371715445418,
|
||
|
|
"grad_norm": 0.9712333083152771,
|
||
|
|
"learning_rate": 1.6836506903848082e-05,
|
||
|
|
"loss": 0.4802510738372803,
|
||
|
|
"mean_token_accuracy": 0.8601484194397926,
|
||
|
|
"num_tokens": 10271784.0,
|
||
|
|
"step": 1810
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4092728039249778,
|
||
|
|
"epoch": 1.554582354197821,
|
||
|
|
"grad_norm": 0.7835445404052734,
|
||
|
|
"learning_rate": 1.6688461151462686e-05,
|
||
|
|
"loss": 0.41991028785705564,
|
||
|
|
"mean_token_accuracy": 0.8713715635240078,
|
||
|
|
"num_tokens": 10328007.0,
|
||
|
|
"step": 1820
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.45985435927286744,
|
||
|
|
"epoch": 1.5631275368511002,
|
||
|
|
"grad_norm": 0.851019561290741,
|
||
|
|
"learning_rate": 1.6540248621273082e-05,
|
||
|
|
"loss": 0.4471141338348389,
|
||
|
|
"mean_token_accuracy": 0.8618485115468502,
|
||
|
|
"num_tokens": 10381848.0,
|
||
|
|
"step": 1830
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.48469415297731755,
|
||
|
|
"epoch": 1.5716727195043794,
|
||
|
|
"grad_norm": 0.9116345047950745,
|
||
|
|
"learning_rate": 1.639188395297668e-05,
|
||
|
|
"loss": 0.4673773288726807,
|
||
|
|
"mean_token_accuracy": 0.8544772244989872,
|
||
|
|
"num_tokens": 10441374.0,
|
||
|
|
"step": 1840
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.46359577621333303,
|
||
|
|
"epoch": 1.5802179021576586,
|
||
|
|
"grad_norm": 0.753823459148407,
|
||
|
|
"learning_rate": 1.624338180129833e-05,
|
||
|
|
"loss": 0.4436906337738037,
|
||
|
|
"mean_token_accuracy": 0.8599751383066178,
|
||
|
|
"num_tokens": 10493832.0,
|
||
|
|
"step": 1850
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4334666928742081,
|
||
|
|
"epoch": 1.588763084810938,
|
||
|
|
"grad_norm": 0.8658061027526855,
|
||
|
|
"learning_rate": 1.6094756834542823e-05,
|
||
|
|
"loss": 0.4242602825164795,
|
||
|
|
"mean_token_accuracy": 0.8660054102540016,
|
||
|
|
"num_tokens": 10554480.0,
|
||
|
|
"step": 1860
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.42020976906642316,
|
||
|
|
"epoch": 1.597308267464217,
|
||
|
|
"grad_norm": 0.6106055378913879,
|
||
|
|
"learning_rate": 1.594602373314599e-05,
|
||
|
|
"loss": 0.4327824115753174,
|
||
|
|
"mean_token_accuracy": 0.8648590184748173,
|
||
|
|
"num_tokens": 10614222.0,
|
||
|
|
"step": 1870
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.43514787331223487,
|
||
|
|
"epoch": 1.6058534501174964,
|
||
|
|
"grad_norm": 0.9191545844078064,
|
||
|
|
"learning_rate": 1.579719718822467e-05,
|
||
|
|
"loss": 0.435378885269165,
|
||
|
|
"mean_token_accuracy": 0.8660413719713688,
|
||
|
|
"num_tokens": 10670517.0,
|
||
|
|
"step": 1880
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.44422514657489953,
|
||
|
|
"epoch": 1.6143986327707753,
|
||
|
|
"grad_norm": 0.6802365183830261,
|
||
|
|
"learning_rate": 1.564829190012561e-05,
|
||
|
|
"loss": 0.4380777359008789,
|
||
|
|
"mean_token_accuracy": 0.8624177664518357,
|
||
|
|
"num_tokens": 10726355.0,
|
||
|
|
"step": 1890
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"entropy": 0.4497006971854717,
|
||
|
|
"epoch": 1.6229438154240547,
|
||
|
|
"grad_norm": 0.9305739998817444,
|
||
|
|
"learning_rate": 1.549932257697339e-05,
|
||
|
|
"loss": 0.45711960792541506,
|
||
|
|
"mean_token_accuracy": 0.858258007466793,
|
||
|
|
"num_tokens": 10784765.0,
|
||
|
|
"step": 1900
|
||
|
|
},
|
||
|
|
{
|
||
|
|
"epoch": 1.6229438154240547,
|
||
|
|
"eval_entropy": 0.1260934477977351,
|
||
|
|
"eval_loss": 0.255153089761734,
|
||
|
|
"eval_mean_token_accuracy": 0.8871387547964914,
|
||
|
|
"eval_num_tokens": 10784765.0,
|
||
|
|
"eval_runtime": 131.2707,
|
||
|
|
"eval_samples_per_second": 25.185,
|
||
|
|
"eval_steps_per_second": 12.592,
|
||
|
|
"step": 1900
|
||
|
|
}
|
||
|
|
],
|
||
|
|
"logging_steps": 10,
|
||
|
|
"max_steps": 3513,
|
||
|
|
"num_input_tokens_seen": 0,
|
||
|
|
"num_train_epochs": 3,
|
||
|
|
"save_steps": 100,
|
||
|
|
"stateful_callbacks": {
|
||
|
|
"TrainerControl": {
|
||
|
|
"args": {
|
||
|
|
"should_epoch_stop": false,
|
||
|
|
"should_evaluate": false,
|
||
|
|
"should_log": false,
|
||
|
|
"should_save": true,
|
||
|
|
"should_training_stop": false
|
||
|
|
},
|
||
|
|
"attributes": {}
|
||
|
|
}
|
||
|
|
},
|
||
|
|
"total_flos": 1.1677655009057587e+17,
|
||
|
|
"train_batch_size": 2,
|
||
|
|
"trial_name": null,
|
||
|
|
"trial_params": null
|
||
|
|
}
|