{
  "best_metric": 0.35982879996299744,
  "best_model_checkpoint": "./exper_batch_16_e4/checkpoint-2500",
  "epoch": 4.0,
  "global_step": 2556,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 0.0001992175273865415,
      "loss": 4.9483,
      "step": 10
    },
    {
      "epoch": 0.03,
      "learning_rate": 0.00019843505477308294,
      "loss": 4.7934,
      "step": 20
    },
    {
      "epoch": 0.05,
      "learning_rate": 0.00019765258215962445,
      "loss": 4.6354,
      "step": 30
    },
    {
      "epoch": 0.06,
      "learning_rate": 0.0001968701095461659,
      "loss": 4.4364,
      "step": 40
    },
    {
      "epoch": 0.08,
      "learning_rate": 0.00019608763693270738,
      "loss": 4.3212,
      "step": 50
    },
    {
      "epoch": 0.09,
      "learning_rate": 0.00019530516431924883,
      "loss": 4.2134,
      "step": 60
    },
    {
      "epoch": 0.11,
      "learning_rate": 0.0001945226917057903,
      "loss": 4.0868,
      "step": 70
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00019374021909233179,
      "loss": 3.992,
      "step": 80
    },
    {
      "epoch": 0.14,
      "learning_rate": 0.00019295774647887326,
      "loss": 3.9904,
      "step": 90
    },
    {
      "epoch": 0.16,
      "learning_rate": 0.00019217527386541472,
      "loss": 3.7606,
      "step": 100
    },
    {
      "epoch": 0.16,
      "eval_accuracy": 0.19892058596761758,
      "eval_loss": 3.783871650695801,
      "eval_runtime": 41.3855,
      "eval_samples_per_second": 62.679,
      "eval_steps_per_second": 7.853,
      "step": 100
    },
    {
      "epoch": 0.17,
      "learning_rate": 0.0001913928012519562,
      "loss": 3.8002,
      "step": 110
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.00019061032863849765,
      "loss": 3.6512,
      "step": 120
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.00018982785602503915,
      "loss": 3.5586,
      "step": 130
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.0001890453834115806,
      "loss": 3.3856,
      "step": 140
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00018826291079812208,
      "loss": 3.4324,
      "step": 150
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.00018748043818466354,
      "loss": 3.3126,
      "step": 160
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.00018669796557120501,
      "loss": 3.1905,
      "step": 170
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.0001859154929577465,
      "loss": 3.1417,
      "step": 180
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.00018513302034428797,
      "loss": 3.0441,
      "step": 190
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.00018435054773082942,
      "loss": 3.1072,
      "step": 200
    },
    {
      "epoch": 0.31,
      "eval_accuracy": 0.3284502698535081,
      "eval_loss": 3.0251049995422363,
      "eval_runtime": 40.3565,
      "eval_samples_per_second": 64.277,
      "eval_steps_per_second": 8.053,
      "step": 200
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.0001835680751173709,
      "loss": 2.8646,
      "step": 210
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00018278560250391235,
      "loss": 2.9163,
      "step": 220
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.00018200312989045386,
      "loss": 2.9385,
      "step": 230
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.0001812206572769953,
      "loss": 2.7414,
      "step": 240
    },
    {
      "epoch": 0.39,
      "learning_rate": 0.0001804381846635368,
      "loss": 2.628,
      "step": 250
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00017965571205007824,
      "loss": 2.7957,
      "step": 260
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00017887323943661972,
      "loss": 2.6525,
      "step": 270
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.0001780907668231612,
      "loss": 2.7478,
      "step": 280
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.00017730829420970268,
      "loss": 2.4893,
      "step": 290
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00017652582159624413,
      "loss": 2.4068,
      "step": 300
    },
    {
      "epoch": 0.47,
      "eval_accuracy": 0.471858134155744,
      "eval_loss": 2.4379851818084717,
      "eval_runtime": 40.2783,
      "eval_samples_per_second": 64.402,
      "eval_steps_per_second": 8.069,
      "step": 300
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.0001757433489827856,
      "loss": 2.5014,
      "step": 310
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0001749608763693271,
      "loss": 2.3865,
      "step": 320
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00017417840375586857,
      "loss": 2.3867,
      "step": 330
    },
    {
      "epoch": 0.53,
      "learning_rate": 0.00017339593114241002,
      "loss": 2.4403,
      "step": 340
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.0001726134585289515,
      "loss": 2.2411,
      "step": 350
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00017183098591549295,
      "loss": 2.0845,
      "step": 360
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.00017104851330203443,
      "loss": 2.1195,
      "step": 370
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.0001702660406885759,
      "loss": 2.0963,
      "step": 380
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.00016948356807511739,
      "loss": 2.1239,
      "step": 390
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00016870109546165884,
      "loss": 2.0881,
      "step": 400
    },
    {
      "epoch": 0.63,
      "eval_accuracy": 0.5412490362374711,
      "eval_loss": 2.048940658569336,
      "eval_runtime": 40.084,
      "eval_samples_per_second": 64.714,
      "eval_steps_per_second": 8.108,
      "step": 400
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.00016791862284820032,
      "loss": 2.0617,
      "step": 410
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.0001671361502347418,
      "loss": 2.0724,
      "step": 420
    },
    {
      "epoch": 0.67,
      "learning_rate": 0.00016635367762128327,
      "loss": 1.8695,
      "step": 430
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.00016557120500782473,
      "loss": 1.8601,
      "step": 440
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.0001647887323943662,
      "loss": 1.8267,
      "step": 450
    },
    {
      "epoch": 0.72,
      "learning_rate": 0.00016400625978090766,
      "loss": 1.9212,
      "step": 460
    },
    {
      "epoch": 0.74,
      "learning_rate": 0.00016322378716744916,
      "loss": 1.881,
      "step": 470
    },
    {
      "epoch": 0.75,
      "learning_rate": 0.00016244131455399061,
      "loss": 1.8141,
      "step": 480
    },
    {
      "epoch": 0.77,
      "learning_rate": 0.0001616588419405321,
      "loss": 1.8193,
      "step": 490
    },
    {
      "epoch": 0.78,
      "learning_rate": 0.00016087636932707354,
      "loss": 1.6817,
      "step": 500
    },
    {
      "epoch": 0.78,
      "eval_accuracy": 0.60254433307633,
      "eval_loss": 1.7968339920043945,
      "eval_runtime": 40.4492,
      "eval_samples_per_second": 64.13,
      "eval_steps_per_second": 8.035,
      "step": 500
    },
    {
      "epoch": 0.8,
      "learning_rate": 0.00016009389671361502,
      "loss": 1.5543,
      "step": 510
    },
    {
      "epoch": 0.81,
      "learning_rate": 0.0001593114241001565,
      "loss": 1.638,
      "step": 520
    },
    {
      "epoch": 0.83,
      "learning_rate": 0.00015852895148669798,
      "loss": 1.6943,
      "step": 530
    },
    {
      "epoch": 0.85,
      "learning_rate": 0.00015774647887323943,
      "loss": 1.5593,
      "step": 540
    },
    {
      "epoch": 0.86,
      "learning_rate": 0.0001569640062597809,
      "loss": 1.5751,
      "step": 550
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.00015618153364632236,
      "loss": 1.4714,
      "step": 560
    },
    {
      "epoch": 0.89,
      "learning_rate": 0.00015539906103286387,
      "loss": 1.3496,
      "step": 570
    },
    {
      "epoch": 0.91,
      "learning_rate": 0.00015461658841940532,
      "loss": 1.458,
      "step": 580
    },
    {
      "epoch": 0.92,
      "learning_rate": 0.0001538341158059468,
      "loss": 1.291,
      "step": 590
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.00015305164319248828,
      "loss": 1.342,
      "step": 600
    },
    {
      "epoch": 0.94,
      "eval_accuracy": 0.6249036237471087,
      "eval_loss": 1.5044325590133667,
      "eval_runtime": 40.4134,
      "eval_samples_per_second": 64.187,
      "eval_steps_per_second": 8.042,
      "step": 600
    },
    {
      "epoch": 0.95,
      "learning_rate": 0.00015226917057902973,
      "loss": 1.519,
      "step": 610
    },
    {
      "epoch": 0.97,
      "learning_rate": 0.0001514866979655712,
      "loss": 1.4489,
      "step": 620
    },
    {
      "epoch": 0.99,
      "learning_rate": 0.0001507042253521127,
      "loss": 1.5582,
      "step": 630
    },
    {
      "epoch": 1.0,
      "learning_rate": 0.00014992175273865417,
      "loss": 1.1882,
      "step": 640
    },
    {
      "epoch": 1.02,
      "learning_rate": 0.00014913928012519562,
      "loss": 1.0538,
      "step": 650
    },
    {
      "epoch": 1.03,
      "learning_rate": 0.0001483568075117371,
      "loss": 1.0577,
      "step": 660
    },
    {
      "epoch": 1.05,
      "learning_rate": 0.00014757433489827858,
      "loss": 1.0974,
      "step": 670
    },
    {
      "epoch": 1.06,
      "learning_rate": 0.00014679186228482006,
      "loss": 0.9575,
      "step": 680
    },
    {
      "epoch": 1.08,
      "learning_rate": 0.0001460093896713615,
      "loss": 1.0651,
      "step": 690
    },
    {
      "epoch": 1.1,
      "learning_rate": 0.00014522691705790299,
      "loss": 0.9343,
      "step": 700
    },
    {
      "epoch": 1.1,
      "eval_accuracy": 0.7131842713955281,
      "eval_loss": 1.1881427764892578,
      "eval_runtime": 40.2414,
      "eval_samples_per_second": 64.461,
      "eval_steps_per_second": 8.076,
      "step": 700
    },
    {
      "epoch": 1.11,
      "learning_rate": 0.00014444444444444444,
      "loss": 0.8771,
      "step": 710
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.00014366197183098594,
      "loss": 0.936,
      "step": 720
    },
    {
      "epoch": 1.14,
      "learning_rate": 0.0001428794992175274,
      "loss": 0.9483,
      "step": 730
    },
    {
      "epoch": 1.16,
      "learning_rate": 0.00014209702660406887,
      "loss": 0.8231,
      "step": 740
    },
    {
      "epoch": 1.17,
      "learning_rate": 0.00014131455399061033,
      "loss": 0.8502,
      "step": 750
    },
    {
      "epoch": 1.19,
      "learning_rate": 0.0001405320813771518,
      "loss": 0.9962,
      "step": 760
    },
    {
      "epoch": 1.21,
      "learning_rate": 0.00013974960876369328,
      "loss": 0.7861,
      "step": 770
    },
    {
      "epoch": 1.22,
      "learning_rate": 0.00013896713615023476,
      "loss": 1.121,
      "step": 780
    },
    {
      "epoch": 1.24,
      "learning_rate": 0.00013818466353677621,
      "loss": 0.8718,
      "step": 790
    },
    {
      "epoch": 1.25,
      "learning_rate": 0.0001374021909233177,
      "loss": 0.9552,
      "step": 800
    },
    {
      "epoch": 1.25,
      "eval_accuracy": 0.7224363916730917,
      "eval_loss": 1.1064492464065552,
      "eval_runtime": 40.4181,
      "eval_samples_per_second": 64.179,
      "eval_steps_per_second": 8.041,
      "step": 800
    },
    {
      "epoch": 1.27,
      "learning_rate": 0.00013661971830985914,
      "loss": 0.999,
      "step": 810
    },
    {
      "epoch": 1.28,
      "learning_rate": 0.00013583724569640065,
      "loss": 0.9067,
      "step": 820
    },
    {
      "epoch": 1.3,
      "learning_rate": 0.0001350547730829421,
      "loss": 0.9772,
      "step": 830
    },
    {
      "epoch": 1.31,
      "learning_rate": 0.00013427230046948358,
      "loss": 0.7953,
      "step": 840
    },
    {
      "epoch": 1.33,
      "learning_rate": 0.00013348982785602503,
      "loss": 0.9132,
      "step": 850
    },
    {
      "epoch": 1.35,
      "learning_rate": 0.0001327073552425665,
      "loss": 0.8452,
      "step": 860
    },
    {
      "epoch": 1.36,
      "learning_rate": 0.000131924882629108,
      "loss": 0.8477,
      "step": 870
    },
    {
      "epoch": 1.38,
      "learning_rate": 0.00013114241001564947,
      "loss": 0.7612,
      "step": 880
    },
    {
      "epoch": 1.39,
      "learning_rate": 0.00013035993740219092,
      "loss": 0.8274,
      "step": 890
    },
    {
      "epoch": 1.41,
      "learning_rate": 0.0001295774647887324,
      "loss": 0.7265,
      "step": 900
    },
    {
      "epoch": 1.41,
      "eval_accuracy": 0.776792598303778,
      "eval_loss": 0.9189362525939941,
      "eval_runtime": 40.6855,
      "eval_samples_per_second": 63.757,
      "eval_steps_per_second": 7.988,
      "step": 900
    },
    {
      "epoch": 1.42,
      "learning_rate": 0.00012879499217527385,
      "loss": 0.7843,
      "step": 910
    },
    {
      "epoch": 1.44,
      "learning_rate": 0.00012801251956181536,
      "loss": 0.6628,
      "step": 920
    },
    {
      "epoch": 1.46,
      "learning_rate": 0.0001272300469483568,
      "loss": 0.7944,
      "step": 930
    },
    {
      "epoch": 1.47,
      "learning_rate": 0.0001264475743348983,
      "loss": 0.8215,
      "step": 940
    },
    {
      "epoch": 1.49,
      "learning_rate": 0.00012566510172143974,
      "loss": 0.7672,
      "step": 950
    },
    {
      "epoch": 1.5,
      "learning_rate": 0.00012488262910798122,
      "loss": 0.7514,
      "step": 960
    },
    {
      "epoch": 1.52,
      "learning_rate": 0.0001241001564945227,
      "loss": 0.5692,
      "step": 970
    },
    {
      "epoch": 1.53,
      "learning_rate": 0.00012331768388106418,
      "loss": 0.6487,
      "step": 980
    },
    {
      "epoch": 1.55,
      "learning_rate": 0.00012253521126760563,
      "loss": 0.7988,
      "step": 990
    },
    {
      "epoch": 1.56,
      "learning_rate": 0.00012175273865414711,
      "loss": 0.6732,
      "step": 1000
    },
    {
      "epoch": 1.56,
      "eval_accuracy": 0.7606013878180417,
      "eval_loss": 0.9226632118225098,
      "eval_runtime": 40.3502,
      "eval_samples_per_second": 64.287,
      "eval_steps_per_second": 8.054,
      "step": 1000
    },
    {
      "epoch": 1.58,
      "learning_rate": 0.00012097026604068857,
      "loss": 0.6441,
      "step": 1010
    },
    {
      "epoch": 1.6,
      "learning_rate": 0.00012018779342723005,
      "loss": 0.6007,
      "step": 1020
    },
    {
      "epoch": 1.61,
      "learning_rate": 0.00011940532081377152,
      "loss": 0.5589,
      "step": 1030
    },
    {
      "epoch": 1.63,
      "learning_rate": 0.000118622848200313,
      "loss": 0.6297,
      "step": 1040
    },
    {
      "epoch": 1.64,
      "learning_rate": 0.00011784037558685446,
      "loss": 0.6216,
      "step": 1050
    },
    {
      "epoch": 1.66,
      "learning_rate": 0.00011705790297339594,
      "loss": 0.5161,
      "step": 1060
    },
    {
      "epoch": 1.67,
      "learning_rate": 0.0001162754303599374,
      "loss": 0.5566,
      "step": 1070
    },
    {
      "epoch": 1.69,
      "learning_rate": 0.00011549295774647888,
      "loss": 0.7403,
      "step": 1080
    },
    {
      "epoch": 1.71,
      "learning_rate": 0.00011471048513302034,
      "loss": 0.6499,
      "step": 1090
    },
    {
      "epoch": 1.72,
      "learning_rate": 0.00011392801251956183,
      "loss": 0.5587,
      "step": 1100
    },
    {
      "epoch": 1.72,
      "eval_accuracy": 0.7902852737085582,
      "eval_loss": 0.7912017703056335,
      "eval_runtime": 40.6432,
      "eval_samples_per_second": 63.824,
      "eval_steps_per_second": 7.996,
      "step": 1100
    },
    {
      "epoch": 1.74,
      "learning_rate": 0.00011314553990610328,
      "loss": 0.6908,
      "step": 1110
    },
    {
      "epoch": 1.75,
      "learning_rate": 0.00011236306729264476,
      "loss": 0.4689,
      "step": 1120
    },
    {
      "epoch": 1.77,
      "learning_rate": 0.00011158059467918622,
      "loss": 0.6043,
      "step": 1130
    },
    {
      "epoch": 1.78,
      "learning_rate": 0.0001107981220657277,
      "loss": 0.507,
      "step": 1140
    },
    {
      "epoch": 1.8,
      "learning_rate": 0.00011001564945226917,
      "loss": 0.4835,
      "step": 1150
    },
    {
      "epoch": 1.82,
      "learning_rate": 0.00010923317683881065,
      "loss": 0.6085,
      "step": 1160
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.00010845070422535213,
      "loss": 0.6501,
      "step": 1170
    },
    {
      "epoch": 1.85,
      "learning_rate": 0.00010766823161189359,
      "loss": 0.3569,
      "step": 1180
    },
    {
      "epoch": 1.86,
      "learning_rate": 0.00010688575899843507,
      "loss": 0.531,
      "step": 1190
    },
    {
      "epoch": 1.88,
      "learning_rate": 0.00010610328638497653,
      "loss": 0.6332,
      "step": 1200
    },
    {
      "epoch": 1.88,
      "eval_accuracy": 0.7945258288357748,
      "eval_loss": 0.76058030128479,
      "eval_runtime": 40.0299,
      "eval_samples_per_second": 64.801,
      "eval_steps_per_second": 8.119,
      "step": 1200
    },
    {
      "epoch": 1.89,
      "learning_rate": 0.00010532081377151801,
      "loss": 0.6965,
      "step": 1210
    },
    {
      "epoch": 1.91,
      "learning_rate": 0.00010453834115805947,
      "loss": 0.598,
      "step": 1220
    },
    {
      "epoch": 1.92,
      "learning_rate": 0.00010375586854460096,
      "loss": 0.719,
      "step": 1230
    },
    {
      "epoch": 1.94,
      "learning_rate": 0.00010297339593114241,
      "loss": 0.5609,
      "step": 1240
    },
    {
      "epoch": 1.96,
      "learning_rate": 0.0001021909233176839,
      "loss": 0.5989,
      "step": 1250
    },
    {
      "epoch": 1.97,
      "learning_rate": 0.00010140845070422535,
      "loss": 0.5327,
      "step": 1260
    },
    {
      "epoch": 1.99,
      "learning_rate": 0.00010062597809076683,
      "loss": 0.5099,
      "step": 1270
    },
    {
      "epoch": 2.0,
      "learning_rate": 9.98435054773083e-05,
      "loss": 0.4551,
      "step": 1280
    },
    {
      "epoch": 2.02,
      "learning_rate": 9.906103286384976e-05,
      "loss": 0.2834,
      "step": 1290
    },
    {
      "epoch": 2.03,
      "learning_rate": 9.827856025039124e-05,
      "loss": 0.3188,
      "step": 1300
    },
    {
      "epoch": 2.03,
      "eval_accuracy": 0.8288357748650732,
      "eval_loss": 0.6534976363182068,
      "eval_runtime": 40.5835,
      "eval_samples_per_second": 63.918,
      "eval_steps_per_second": 8.008,
      "step": 1300
    },
    {
      "epoch": 2.05,
      "learning_rate": 9.749608763693271e-05,
      "loss": 0.2455,
      "step": 1310
    },
    {
      "epoch": 2.07,
      "learning_rate": 9.671361502347419e-05,
      "loss": 0.3339,
      "step": 1320
    },
    {
      "epoch": 2.08,
      "learning_rate": 9.593114241001565e-05,
      "loss": 0.2413,
      "step": 1330
    },
    {
      "epoch": 2.1,
      "learning_rate": 9.514866979655712e-05,
      "loss": 0.3398,
      "step": 1340
    },
    {
      "epoch": 2.11,
      "learning_rate": 9.43661971830986e-05,
      "loss": 0.2706,
      "step": 1350
    },
    {
      "epoch": 2.13,
      "learning_rate": 9.358372456964006e-05,
      "loss": 0.3681,
      "step": 1360
    },
    {
      "epoch": 2.14,
      "learning_rate": 9.280125195618154e-05,
      "loss": 0.2871,
      "step": 1370
    },
    {
      "epoch": 2.16,
      "learning_rate": 9.2018779342723e-05,
      "loss": 0.2859,
      "step": 1380
    },
    {
      "epoch": 2.18,
      "learning_rate": 9.123630672926447e-05,
      "loss": 0.1933,
      "step": 1390
    },
    {
      "epoch": 2.19,
      "learning_rate": 9.045383411580595e-05,
      "loss": 0.3079,
      "step": 1400
    },
    {
      "epoch": 2.19,
      "eval_accuracy": 0.8577486507324595,
      "eval_loss": 0.5686159133911133,
      "eval_runtime": 40.5767,
      "eval_samples_per_second": 63.928,
      "eval_steps_per_second": 8.01,
      "step": 1400
    },
    {
      "epoch": 2.21,
      "learning_rate": 8.967136150234741e-05,
      "loss": 0.222,
      "step": 1410
    },
    {
      "epoch": 2.22,
      "learning_rate": 8.888888888888889e-05,
      "loss": 0.326,
      "step": 1420
    },
    {
      "epoch": 2.24,
      "learning_rate": 8.810641627543036e-05,
      "loss": 0.3696,
      "step": 1430
    },
    {
      "epoch": 2.25,
      "learning_rate": 8.732394366197182e-05,
      "loss": 0.2816,
      "step": 1440
    },
    {
      "epoch": 2.27,
      "learning_rate": 8.65414710485133e-05,
      "loss": 0.2642,
      "step": 1450
    },
    {
      "epoch": 2.28,
      "learning_rate": 8.575899843505478e-05,
      "loss": 0.26,
      "step": 1460
    },
    {
      "epoch": 2.3,
      "learning_rate": 8.497652582159625e-05,
      "loss": 0.2538,
      "step": 1470
    },
    {
      "epoch": 2.32,
      "learning_rate": 8.419405320813773e-05,
      "loss": 0.305,
      "step": 1480
    },
    {
      "epoch": 2.33,
      "learning_rate": 8.341158059467919e-05,
      "loss": 0.3137,
      "step": 1490
    },
    {
      "epoch": 2.35,
      "learning_rate": 8.262910798122067e-05,
      "loss": 0.2518,
      "step": 1500
    },
    {
      "epoch": 2.35,
      "eval_accuracy": 0.8577486507324595,
      "eval_loss": 0.5517102479934692,
      "eval_runtime": 40.6225,
      "eval_samples_per_second": 63.856,
      "eval_steps_per_second": 8.0,
      "step": 1500
    },
    {
      "epoch": 2.36,
      "learning_rate": 8.184663536776213e-05,
      "loss": 0.2775,
      "step": 1510
    },
    {
      "epoch": 2.38,
      "learning_rate": 8.106416275430361e-05,
      "loss": 0.2882,
      "step": 1520
    },
    {
      "epoch": 2.39,
      "learning_rate": 8.028169014084508e-05,
      "loss": 0.2506,
      "step": 1530
    },
    {
      "epoch": 2.41,
      "learning_rate": 7.949921752738654e-05,
      "loss": 0.2111,
      "step": 1540
    },
    {
      "epoch": 2.43,
      "learning_rate": 7.871674491392802e-05,
      "loss": 0.2025,
      "step": 1550
    },
    {
      "epoch": 2.44,
      "learning_rate": 7.793427230046949e-05,
      "loss": 0.2137,
      "step": 1560
    },
    {
      "epoch": 2.46,
      "learning_rate": 7.715179968701097e-05,
      "loss": 0.1643,
      "step": 1570
    },
    {
      "epoch": 2.47,
      "learning_rate": 7.636932707355243e-05,
      "loss": 0.1827,
      "step": 1580
    },
    {
      "epoch": 2.49,
      "learning_rate": 7.55868544600939e-05,
      "loss": 0.1774,
      "step": 1590
    },
    {
      "epoch": 2.5,
      "learning_rate": 7.480438184663538e-05,
      "loss": 0.2,
      "step": 1600
    },
    {
      "epoch": 2.5,
      "eval_accuracy": 0.8631457208943716,
      "eval_loss": 0.5276668667793274,
      "eval_runtime": 40.6494,
      "eval_samples_per_second": 63.814,
      "eval_steps_per_second": 7.995,
      "step": 1600
    },
    {
      "epoch": 2.52,
      "learning_rate": 7.402190923317684e-05,
      "loss": 0.1489,
      "step": 1610
    },
    {
      "epoch": 2.54,
      "learning_rate": 7.323943661971832e-05,
      "loss": 0.1583,
      "step": 1620
    },
    {
      "epoch": 2.55,
      "learning_rate": 7.245696400625979e-05,
      "loss": 0.194,
      "step": 1630
    },
    {
      "epoch": 2.57,
      "learning_rate": 7.167449139280125e-05,
      "loss": 0.3158,
      "step": 1640
    },
    {
      "epoch": 2.58,
      "learning_rate": 7.089201877934273e-05,
      "loss": 0.2098,
      "step": 1650
    },
    {
      "epoch": 2.6,
      "learning_rate": 7.01095461658842e-05,
      "loss": 0.2711,
      "step": 1660
    },
    {
      "epoch": 2.61,
      "learning_rate": 6.932707355242567e-05,
      "loss": 0.2274,
      "step": 1670
    },
    {
      "epoch": 2.63,
      "learning_rate": 6.854460093896714e-05,
      "loss": 0.1626,
      "step": 1680
    },
    {
      "epoch": 2.64,
      "learning_rate": 6.77621283255086e-05,
      "loss": 0.1899,
      "step": 1690
    },
    {
      "epoch": 2.66,
      "learning_rate": 6.697965571205008e-05,
      "loss": 0.2032,
      "step": 1700
    },
    {
      "epoch": 2.66,
      "eval_accuracy": 0.8700848111025443,
      "eval_loss": 0.4840567409992218,
      "eval_runtime": 40.6495,
      "eval_samples_per_second": 63.814,
      "eval_steps_per_second": 7.995,
      "step": 1700
    },
    {
      "epoch": 2.68,
      "learning_rate": 6.619718309859155e-05,
      "loss": 0.1504,
      "step": 1710
    },
    {
      "epoch": 2.69,
      "learning_rate": 6.541471048513303e-05,
      "loss": 0.2503,
      "step": 1720
    },
    {
      "epoch": 2.71,
      "learning_rate": 6.463223787167449e-05,
      "loss": 0.2004,
      "step": 1730
    },
    {
      "epoch": 2.72,
      "learning_rate": 6.384976525821597e-05,
      "loss": 0.2705,
      "step": 1740
    },
    {
      "epoch": 2.74,
      "learning_rate": 6.306729264475744e-05,
      "loss": 0.1809,
      "step": 1750
    },
    {
      "epoch": 2.75,
      "learning_rate": 6.22848200312989e-05,
      "loss": 0.2257,
      "step": 1760
    },
    {
      "epoch": 2.77,
      "learning_rate": 6.150234741784038e-05,
      "loss": 0.1847,
      "step": 1770
    },
    {
      "epoch": 2.79,
      "learning_rate": 6.071987480438185e-05,
      "loss": 0.1488,
      "step": 1780
    },
    {
      "epoch": 2.8,
      "learning_rate": 5.993740219092332e-05,
      "loss": 0.2469,
      "step": 1790
    },
    {
      "epoch": 2.82,
      "learning_rate": 5.915492957746479e-05,
      "loss": 0.1555,
      "step": 1800
    },
    {
      "epoch": 2.82,
      "eval_accuracy": 0.8793369313801079,
      "eval_loss": 0.4578264057636261,
      "eval_runtime": 40.536,
      "eval_samples_per_second": 63.992,
      "eval_steps_per_second": 8.018,
      "step": 1800
    },
    {
      "epoch": 2.83,
      "learning_rate": 5.837245696400626e-05,
      "loss": 0.2625,
      "step": 1810
    },
    {
      "epoch": 2.85,
      "learning_rate": 5.7589984350547735e-05,
      "loss": 0.1896,
      "step": 1820
    },
    {
      "epoch": 2.86,
      "learning_rate": 5.68075117370892e-05,
      "loss": 0.1443,
      "step": 1830
    },
    {
      "epoch": 2.88,
      "learning_rate": 5.602503912363067e-05,
      "loss": 0.205,
      "step": 1840
    },
    {
      "epoch": 2.9,
      "learning_rate": 5.5242566510172144e-05,
      "loss": 0.135,
      "step": 1850
    },
    {
      "epoch": 2.91,
      "learning_rate": 5.4460093896713616e-05,
      "loss": 0.1833,
      "step": 1860
    },
    {
      "epoch": 2.93,
      "learning_rate": 5.367762128325509e-05,
      "loss": 0.1534,
      "step": 1870
    },
    {
      "epoch": 2.94,
      "learning_rate": 5.2895148669796554e-05,
      "loss": 0.2702,
      "step": 1880
    },
    {
      "epoch": 2.96,
      "learning_rate": 5.2112676056338026e-05,
      "loss": 0.2304,
      "step": 1890
    },
    {
      "epoch": 2.97,
      "learning_rate": 5.13302034428795e-05,
      "loss": 0.145,
      "step": 1900
    },
    {
      "epoch": 2.97,
      "eval_accuracy": 0.8754818812644565,
      "eval_loss": 0.4465731978416443,
      "eval_runtime": 39.9438,
      "eval_samples_per_second": 64.941,
      "eval_steps_per_second": 8.136,
      "step": 1900
    },
    {
      "epoch": 2.99,
      "learning_rate": 5.054773082942097e-05,
      "loss": 0.1804,
      "step": 1910
    },
    {
      "epoch": 3.0,
      "learning_rate": 4.976525821596245e-05,
      "loss": 0.1421,
      "step": 1920
    },
    {
      "epoch": 3.02,
      "learning_rate": 4.8982785602503914e-05,
      "loss": 0.0721,
      "step": 1930
    },
    {
      "epoch": 3.04,
      "learning_rate": 4.8200312989045386e-05,
      "loss": 0.0716,
      "step": 1940
    },
    {
      "epoch": 3.05,
      "learning_rate": 4.741784037558686e-05,
      "loss": 0.1382,
      "step": 1950
    },
    {
      "epoch": 3.07,
      "learning_rate": 4.663536776212833e-05,
      "loss": 0.103,
      "step": 1960
    },
    {
      "epoch": 3.08,
      "learning_rate": 4.58528951486698e-05,
      "loss": 0.0979,
      "step": 1970
    },
    {
      "epoch": 3.1,
      "learning_rate": 4.507042253521127e-05,
      "loss": 0.0934,
      "step": 1980
    },
    {
      "epoch": 3.11,
      "learning_rate": 4.428794992175274e-05,
      "loss": 0.0561,
      "step": 1990
    },
    {
      "epoch": 3.13,
      "learning_rate": 4.350547730829421e-05,
      "loss": 0.0985,
      "step": 2000
    },
    {
      "epoch": 3.13,
      "eval_accuracy": 0.8866615265998458,
      "eval_loss": 0.4248804450035095,
      "eval_runtime": 39.986,
      "eval_samples_per_second": 64.873,
      "eval_steps_per_second": 8.128,
      "step": 2000
    },
    {
      "epoch": 3.15,
      "learning_rate": 4.2723004694835684e-05,
      "loss": 0.0915,
      "step": 2010
    },
    {
      "epoch": 3.16,
      "learning_rate": 4.1940532081377156e-05,
      "loss": 0.1015,
      "step": 2020
    },
    {
      "epoch": 3.18,
      "learning_rate": 4.115805946791863e-05,
      "loss": 0.1064,
      "step": 2030
    },
    {
      "epoch": 3.19,
      "learning_rate": 4.037558685446009e-05,
      "loss": 0.0545,
      "step": 2040
    },
    {
      "epoch": 3.21,
      "learning_rate": 3.9593114241001565e-05,
      "loss": 0.0653,
      "step": 2050
    },
    {
      "epoch": 3.22,
      "learning_rate": 3.881064162754304e-05,
      "loss": 0.0589,
      "step": 2060
    },
    {
      "epoch": 3.24,
      "learning_rate": 3.802816901408451e-05,
      "loss": 0.073,
      "step": 2070
    },
    {
      "epoch": 3.26,
      "learning_rate": 3.724569640062598e-05,
      "loss": 0.0783,
      "step": 2080
    },
    {
      "epoch": 3.27,
      "learning_rate": 3.646322378716745e-05,
      "loss": 0.0746,
      "step": 2090
    },
    {
      "epoch": 3.29,
      "learning_rate": 3.568075117370892e-05,
      "loss": 0.0955,
      "step": 2100
    },
    {
      "epoch": 3.29,
      "eval_accuracy": 0.8932151117964534,
      "eval_loss": 0.3977428078651428,
      "eval_runtime": 40.7031,
      "eval_samples_per_second": 63.73,
      "eval_steps_per_second": 7.985,
      "step": 2100
    },
    {
      "epoch": 3.3,
      "learning_rate": 3.489827856025039e-05,
      "loss": 0.0682,
      "step": 2110
    },
    {
      "epoch": 3.32,
      "learning_rate": 3.411580594679186e-05,
      "loss": 0.05,
      "step": 2120
    },
    {
      "epoch": 3.33,
      "learning_rate": 3.3333333333333335e-05,
      "loss": 0.0565,
      "step": 2130
    },
    {
      "epoch": 3.35,
      "learning_rate": 3.255086071987481e-05,
      "loss": 0.1286,
      "step": 2140
    },
    {
      "epoch": 3.36,
      "learning_rate": 3.176838810641627e-05,
      "loss": 0.0797,
      "step": 2150
    },
    {
      "epoch": 3.38,
      "learning_rate": 3.0985915492957744e-05,
      "loss": 0.0585,
      "step": 2160
    },
    {
      "epoch": 3.4,
      "learning_rate": 3.0203442879499216e-05,
      "loss": 0.0914,
      "step": 2170
    },
    {
      "epoch": 3.41,
      "learning_rate": 2.9420970266040692e-05,
      "loss": 0.1152,
      "step": 2180
    },
    {
      "epoch": 3.43,
      "learning_rate": 2.8638497652582164e-05,
      "loss": 0.035,
      "step": 2190
    },
    {
      "epoch": 3.44,
      "learning_rate": 2.7856025039123636e-05,
      "loss": 0.0438,
      "step": 2200
    },
    {
      "epoch": 3.44,
      "eval_accuracy": 0.9036237471087124,
      "eval_loss": 0.37847575545310974,
      "eval_runtime": 40.8463,
      "eval_samples_per_second": 63.506,
      "eval_steps_per_second": 7.957,
      "step": 2200
    },
    {
      "epoch": 3.46,
      "learning_rate": 2.7073552425665105e-05,
      "loss": 0.0348,
      "step": 2210
    },
    {
      "epoch": 3.47,
      "learning_rate": 2.6291079812206577e-05,
      "loss": 0.0641,
      "step": 2220
    },
    {
      "epoch": 3.49,
      "learning_rate": 2.5508607198748045e-05,
      "loss": 0.062,
      "step": 2230
    },
    {
      "epoch": 3.51,
      "learning_rate": 2.4726134585289514e-05,
      "loss": 0.0845,
      "step": 2240
    },
    {
      "epoch": 3.52,
      "learning_rate": 2.3943661971830986e-05,
      "loss": 0.05,
      "step": 2250
    },
    {
      "epoch": 3.54,
      "learning_rate": 2.3161189358372458e-05,
      "loss": 0.0964,
      "step": 2260
    },
    {
      "epoch": 3.55,
      "learning_rate": 2.237871674491393e-05,
      "loss": 0.0641,
      "step": 2270
    },
    {
      "epoch": 3.57,
      "learning_rate": 2.1596244131455402e-05,
      "loss": 0.0484,
      "step": 2280
    },
    {
      "epoch": 3.58,
      "learning_rate": 2.081377151799687e-05,
      "loss": 0.0393,
      "step": 2290
    },
    {
      "epoch": 3.6,
      "learning_rate": 2.0031298904538343e-05,
      "loss": 0.0589,
      "step": 2300
    },
    {
      "epoch": 3.6,
      "eval_accuracy": 0.9016962220508866,
      "eval_loss": 0.37170541286468506,
      "eval_runtime": 40.4294,
      "eval_samples_per_second": 64.161,
      "eval_steps_per_second": 8.039,
      "step": 2300
    },
    {
      "epoch": 3.62,
      "learning_rate": 1.9248826291079812e-05,
      "loss": 0.0749,
      "step": 2310
    },
    {
      "epoch": 3.63,
      "learning_rate": 1.8466353677621284e-05,
      "loss": 0.0486,
      "step": 2320
    },
    {
      "epoch": 3.65,
      "learning_rate": 1.7683881064162756e-05,
      "loss": 0.0492,
      "step": 2330
    },
    {
      "epoch": 3.66,
      "learning_rate": 1.6901408450704224e-05,
      "loss": 0.0822,
      "step": 2340
    },
    {
      "epoch": 3.68,
      "learning_rate": 1.6118935837245697e-05,
      "loss": 0.0316,
      "step": 2350
    },
    {
      "epoch": 3.69,
      "learning_rate": 1.533646322378717e-05,
      "loss": 0.0388,
      "step": 2360
    },
    {
      "epoch": 3.71,
      "learning_rate": 1.4553990610328639e-05,
      "loss": 0.033,
      "step": 2370
    },
    {
      "epoch": 3.72,
      "learning_rate": 1.377151799687011e-05,
      "loss": 0.0351,
      "step": 2380
    },
    {
      "epoch": 3.74,
      "learning_rate": 1.298904538341158e-05,
      "loss": 0.0724,
      "step": 2390
    },
    {
      "epoch": 3.76,
      "learning_rate": 1.2206572769953052e-05,
      "loss": 0.0709,
      "step": 2400
    },
    {
      "epoch": 3.76,
      "eval_accuracy": 0.905165767154973,
      "eval_loss": 0.3609260618686676,
      "eval_runtime": 40.4534,
      "eval_samples_per_second": 64.123,
      "eval_steps_per_second": 8.034,
      "step": 2400
    },
    {
      "epoch": 3.77,
      "learning_rate": 1.1424100156494522e-05,
      "loss": 0.0492,
      "step": 2410
    },
    {
      "epoch": 3.79,
      "learning_rate": 1.0641627543035994e-05,
      "loss": 0.0414,
      "step": 2420
    },
    {
      "epoch": 3.8,
      "learning_rate": 9.859154929577465e-06,
      "loss": 0.0646,
      "step": 2430
    },
    {
      "epoch": 3.82,
      "learning_rate": 9.076682316118937e-06,
      "loss": 0.0489,
      "step": 2440
    },
    {
      "epoch": 3.83,
      "learning_rate": 8.294209702660407e-06,
      "loss": 0.0339,
      "step": 2450
    },
    {
      "epoch": 3.85,
      "learning_rate": 7.511737089201878e-06,
      "loss": 0.0662,
      "step": 2460
    },
    {
      "epoch": 3.87,
      "learning_rate": 6.7292644757433494e-06,
      "loss": 0.0369,
      "step": 2470
    },
    {
      "epoch": 3.88,
      "learning_rate": 5.94679186228482e-06,
      "loss": 0.1006,
      "step": 2480
    },
    {
      "epoch": 3.9,
      "learning_rate": 5.164319248826292e-06,
      "loss": 0.0295,
      "step": 2490
    },
    {
      "epoch": 3.91,
      "learning_rate": 4.381846635367762e-06,
      "loss": 0.0706,
      "step": 2500
    },
    {
      "epoch": 3.91,
      "eval_accuracy": 0.9059367771781033,
      "eval_loss": 0.35982879996299744,
      "eval_runtime": 40.5931,
      "eval_samples_per_second": 63.903,
      "eval_steps_per_second": 8.006,
      "step": 2500
    },
    {
      "epoch": 3.93,
      "learning_rate": 3.5993740219092334e-06,
      "loss": 0.0756,
      "step": 2510
    },
    {
      "epoch": 3.94,
      "learning_rate": 2.8169014084507042e-06,
      "loss": 0.0523,
      "step": 2520
    },
    {
      "epoch": 3.96,
      "learning_rate": 2.0344287949921754e-06,
      "loss": 0.0368,
      "step": 2530
    },
    {
      "epoch": 3.97,
      "learning_rate": 1.2519561815336464e-06,
      "loss": 0.0417,
      "step": 2540
    },
    {
      "epoch": 3.99,
      "learning_rate": 4.694835680751174e-07,
      "loss": 0.0598,
      "step": 2550
    },
    {
      "epoch": 4.0,
      "step": 2556,
      "total_flos": 3.1689424899978854e+18,
      "train_loss": 0.9094190793184794,
      "train_runtime": 2613.3405,
      "train_samples_per_second": 15.628,
      "train_steps_per_second": 0.978
    }
  ],
  "max_steps": 2556,
  "num_train_epochs": 4,
  "total_flos": 3.1689424899978854e+18,
  "trial_name": null,
  "trial_params": null
}