{
  "best_metric": 0.5178616046905518,
  "best_model_checkpoint": "/data/jcanete/all_results/xnli/albeto_large/epochs_4_bs_16_lr_5e-6/checkpoint-48000",
  "epoch": 4.0,
  "global_step": 98176,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 4.974790172750978e-06,
      "loss": 1.0068,
      "step": 500
    },
    {
      "epoch": 0.04,
      "learning_rate": 4.9493766297262065e-06,
      "loss": 0.8946,
      "step": 1000
    },
    {
      "epoch": 0.06,
      "learning_rate": 4.924268660365059e-06,
      "loss": 0.8615,
      "step": 1500
    },
    {
      "epoch": 0.08,
      "learning_rate": 4.899007904172099e-06,
      "loss": 1.0693,
      "step": 2000
    },
    {
      "epoch": 0.08,
      "eval_accuracy": 0.5301204819277109,
      "eval_loss": 0.9672217965126038,
      "eval_runtime": 3.1152,
      "eval_samples_per_second": 799.309,
      "eval_steps_per_second": 50.077,
      "step": 2000
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.873594361147327e-06,
      "loss": 0.9457,
      "step": 2500
    },
    {
      "epoch": 0.12,
      "learning_rate": 4.848129889178619e-06,
      "loss": 0.8099,
      "step": 3000
    },
    {
      "epoch": 0.14,
      "learning_rate": 4.822665417209909e-06,
      "loss": 0.7534,
      "step": 3500
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.7972009452412e-06,
      "loss": 0.7074,
      "step": 4000
    },
    {
      "epoch": 0.16,
      "eval_accuracy": 0.6859437751004016,
      "eval_loss": 0.7257041931152344,
      "eval_runtime": 3.1181,
      "eval_samples_per_second": 798.568,
      "eval_steps_per_second": 50.031,
      "step": 4000
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.7717364732724905e-06,
      "loss": 0.706,
      "step": 4500
    },
    {
      "epoch": 0.2,
      "learning_rate": 4.746272001303782e-06,
      "loss": 0.6929,
      "step": 5000
    },
    {
      "epoch": 0.22,
      "learning_rate": 4.720807529335072e-06,
      "loss": 0.6786,
      "step": 5500
    },
    {
      "epoch": 0.24,
      "learning_rate": 4.6953430573663625e-06,
      "loss": 0.6533,
      "step": 6000
    },
    {
      "epoch": 0.24,
      "eval_accuracy": 0.7156626506024096,
      "eval_loss": 0.7040365934371948,
      "eval_runtime": 3.106,
      "eval_samples_per_second": 801.678,
      "eval_steps_per_second": 50.226,
      "step": 6000
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.669878585397654e-06,
      "loss": 0.7996,
      "step": 6500
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.644414113428944e-06,
      "loss": 0.7019,
      "step": 7000
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.618949641460235e-06,
      "loss": 0.6643,
      "step": 7500
    },
    {
      "epoch": 0.33,
      "learning_rate": 4.593485169491526e-06,
      "loss": 0.6375,
      "step": 8000
    },
    {
      "epoch": 0.33,
      "eval_accuracy": 0.7421686746987952,
      "eval_loss": 0.6504086256027222,
      "eval_runtime": 3.105,
      "eval_samples_per_second": 801.928,
      "eval_steps_per_second": 50.241,
      "step": 8000
    },
    {
      "epoch": 0.35,
      "learning_rate": 4.568020697522816e-06,
      "loss": 0.6397,
      "step": 8500
    },
    {
      "epoch": 0.37,
      "learning_rate": 4.542556225554107e-06,
      "loss": 0.6264,
      "step": 9000
    },
    {
      "epoch": 0.39,
      "learning_rate": 4.517091753585398e-06,
      "loss": 0.6288,
      "step": 9500
    },
    {
      "epoch": 0.41,
      "learning_rate": 4.491627281616689e-06,
      "loss": 0.6158,
      "step": 10000
    },
    {
      "epoch": 0.41,
      "eval_accuracy": 0.7542168674698795,
      "eval_loss": 0.6070215702056885,
      "eval_runtime": 3.1074,
      "eval_samples_per_second": 801.321,
      "eval_steps_per_second": 50.203,
      "step": 10000
    },
    {
      "epoch": 0.43,
      "learning_rate": 4.46616280964798e-06,
      "loss": 0.6155,
      "step": 10500
    },
    {
      "epoch": 0.45,
      "learning_rate": 4.44069833767927e-06,
      "loss": 0.6139,
      "step": 11000
    },
    {
      "epoch": 0.47,
      "learning_rate": 4.415233865710561e-06,
      "loss": 0.6096,
      "step": 11500
    },
    {
      "epoch": 0.49,
      "learning_rate": 4.389769393741852e-06,
      "loss": 0.5992,
      "step": 12000
    },
    {
      "epoch": 0.49,
      "eval_accuracy": 0.7674698795180723,
      "eval_loss": 0.5774316191673279,
      "eval_runtime": 3.0887,
      "eval_samples_per_second": 806.17,
      "eval_steps_per_second": 50.507,
      "step": 12000
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.364304921773142e-06,
      "loss": 0.6005,
      "step": 12500
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.3388404498044335e-06,
      "loss": 0.6118,
      "step": 13000
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.313375977835724e-06,
      "loss": 0.6063,
      "step": 13500
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.287911505867014e-06,
      "loss": 0.5997,
      "step": 14000
    },
    {
      "epoch": 0.57,
      "eval_accuracy": 0.7714859437751004,
      "eval_loss": 0.5825369954109192,
      "eval_runtime": 3.1126,
      "eval_samples_per_second": 799.965,
      "eval_steps_per_second": 50.118,
      "step": 14000
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.2624470338983054e-06,
      "loss": 0.5779,
      "step": 14500
    },
    {
      "epoch": 0.61,
      "learning_rate": 4.236982561929597e-06,
      "loss": 0.6107,
      "step": 15000
    },
    {
      "epoch": 0.63,
      "learning_rate": 4.211518089960887e-06,
      "loss": 0.5816,
      "step": 15500
    },
    {
      "epoch": 0.65,
      "learning_rate": 4.186053617992178e-06,
      "loss": 0.592,
      "step": 16000
    },
    {
      "epoch": 0.65,
      "eval_accuracy": 0.7742971887550201,
      "eval_loss": 0.5646522641181946,
      "eval_runtime": 3.1034,
      "eval_samples_per_second": 802.352,
      "eval_steps_per_second": 50.268,
      "step": 16000
    },
    {
      "epoch": 0.67,
      "learning_rate": 4.1605891460234686e-06,
      "loss": 0.5846,
      "step": 16500
    },
    {
      "epoch": 0.69,
      "learning_rate": 4.135124674054759e-06,
      "loss": 0.5791,
      "step": 17000
    },
    {
      "epoch": 0.71,
      "learning_rate": 4.10966020208605e-06,
      "loss": 0.595,
      "step": 17500
    },
    {
      "epoch": 0.73,
      "learning_rate": 4.0841957301173405e-06,
      "loss": 0.5682,
      "step": 18000
    },
    {
      "epoch": 0.73,
      "eval_accuracy": 0.7746987951807229,
      "eval_loss": 0.5749292969703674,
      "eval_runtime": 3.0899,
      "eval_samples_per_second": 805.855,
      "eval_steps_per_second": 50.487,
      "step": 18000
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.058731258148632e-06,
      "loss": 0.5634,
      "step": 18500
    },
    {
      "epoch": 0.77,
      "learning_rate": 4.033266786179922e-06,
      "loss": 0.5767,
      "step": 19000
    },
    {
      "epoch": 0.79,
      "learning_rate": 4.0078023142112124e-06,
      "loss": 0.5757,
      "step": 19500
    },
    {
      "epoch": 0.81,
      "learning_rate": 3.982337842242504e-06,
      "loss": 0.5549,
      "step": 20000
    },
    {
      "epoch": 0.81,
      "eval_accuracy": 0.7738955823293173,
      "eval_loss": 0.5731613039970398,
      "eval_runtime": 3.1136,
      "eval_samples_per_second": 799.722,
      "eval_steps_per_second": 50.103,
      "step": 20000
    },
    {
      "epoch": 0.84,
      "learning_rate": 3.956873370273794e-06,
      "loss": 0.5704,
      "step": 20500
    },
    {
      "epoch": 0.86,
      "learning_rate": 3.931408898305085e-06,
      "loss": 0.5668,
      "step": 21000
    },
    {
      "epoch": 0.88,
      "learning_rate": 3.9059444263363764e-06,
      "loss": 0.5523,
      "step": 21500
    },
    {
      "epoch": 0.9,
      "learning_rate": 3.880479954367666e-06,
      "loss": 0.5651,
      "step": 22000
    },
    {
      "epoch": 0.9,
      "eval_accuracy": 0.7887550200803213,
      "eval_loss": 0.5458173751831055,
      "eval_runtime": 3.1133,
      "eval_samples_per_second": 799.785,
      "eval_steps_per_second": 50.107,
      "step": 22000
    },
    {
      "epoch": 0.92,
      "learning_rate": 3.855015482398957e-06,
      "loss": 0.5582,
      "step": 22500
    },
    {
      "epoch": 0.94,
      "learning_rate": 3.829551010430248e-06,
      "loss": 0.5567,
      "step": 23000
    },
    {
      "epoch": 0.96,
      "learning_rate": 3.8040865384615387e-06,
      "loss": 0.5505,
      "step": 23500
    },
    {
      "epoch": 0.98,
      "learning_rate": 3.7786220664928295e-06,
      "loss": 0.561,
      "step": 24000
    },
    {
      "epoch": 0.98,
      "eval_accuracy": 0.7859437751004016,
      "eval_loss": 0.5578823089599609,
      "eval_runtime": 3.1003,
      "eval_samples_per_second": 803.157,
      "eval_steps_per_second": 50.318,
      "step": 24000
    },
    {
      "epoch": 1.0,
      "learning_rate": 3.75315759452412e-06,
      "loss": 0.559,
      "step": 24500
    },
    {
      "epoch": 1.02,
      "learning_rate": 3.7276931225554107e-06,
      "loss": 0.5078,
      "step": 25000
    },
    {
      "epoch": 1.04,
      "learning_rate": 3.702228650586702e-06,
      "loss": 0.5162,
      "step": 25500
    },
    {
      "epoch": 1.06,
      "learning_rate": 3.6767641786179927e-06,
      "loss": 0.5033,
      "step": 26000
    },
    {
      "epoch": 1.06,
      "eval_accuracy": 0.7803212851405622,
      "eval_loss": 0.5404613614082336,
      "eval_runtime": 3.1086,
      "eval_samples_per_second": 801.01,
      "eval_steps_per_second": 50.184,
      "step": 26000
    },
    {
      "epoch": 1.08,
      "learning_rate": 3.6512997066492834e-06,
      "loss": 0.5035,
      "step": 26500
    },
    {
      "epoch": 1.1,
      "learning_rate": 3.6258352346805742e-06,
      "loss": 0.4876,
      "step": 27000
    },
    {
      "epoch": 1.12,
      "learning_rate": 3.6003707627118646e-06,
      "loss": 0.5037,
      "step": 27500
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.5749062907431554e-06,
      "loss": 0.507,
      "step": 28000
    },
    {
      "epoch": 1.14,
      "eval_accuracy": 0.7867469879518072,
      "eval_loss": 0.5373629331588745,
      "eval_runtime": 3.1109,
      "eval_samples_per_second": 800.404,
      "eval_steps_per_second": 50.146,
      "step": 28000
    },
    {
      "epoch": 1.16,
      "learning_rate": 3.549441818774446e-06,
      "loss": 0.5227,
      "step": 28500
    },
    {
      "epoch": 1.18,
      "learning_rate": 3.523977346805737e-06,
      "loss": 0.5025,
      "step": 29000
    },
    {
      "epoch": 1.2,
      "learning_rate": 3.4985128748370277e-06,
      "loss": 0.5202,
      "step": 29500
    },
    {
      "epoch": 1.22,
      "learning_rate": 3.4730484028683185e-06,
      "loss": 0.5027,
      "step": 30000
    },
    {
      "epoch": 1.22,
      "eval_accuracy": 0.7843373493975904,
      "eval_loss": 0.5514963865280151,
      "eval_runtime": 3.114,
      "eval_samples_per_second": 799.625,
      "eval_steps_per_second": 50.097,
      "step": 30000
    },
    {
      "epoch": 1.24,
      "learning_rate": 3.4476348598435467e-06,
      "loss": 0.5052,
      "step": 30500
    },
    {
      "epoch": 1.26,
      "learning_rate": 3.4221703878748375e-06,
      "loss": 0.4932,
      "step": 31000
    },
    {
      "epoch": 1.28,
      "learning_rate": 3.3967059159061283e-06,
      "loss": 0.4916,
      "step": 31500
    },
    {
      "epoch": 1.3,
      "learning_rate": 3.3712414439374187e-06,
      "loss": 0.5021,
      "step": 32000
    },
    {
      "epoch": 1.3,
      "eval_accuracy": 0.793574297188755,
      "eval_loss": 0.5289486050605774,
      "eval_runtime": 3.1115,
      "eval_samples_per_second": 800.255,
      "eval_steps_per_second": 50.136,
      "step": 32000
    },
    {
      "epoch": 1.32,
      "learning_rate": 3.3457769719687095e-06,
      "loss": 0.501,
      "step": 32500
    },
    {
      "epoch": 1.34,
      "learning_rate": 3.3203125000000002e-06,
      "loss": 0.4921,
      "step": 33000
    },
    {
      "epoch": 1.36,
      "learning_rate": 3.294848028031291e-06,
      "loss": 0.5051,
      "step": 33500
    },
    {
      "epoch": 1.39,
      "learning_rate": 3.269383556062582e-06,
      "loss": 0.4895,
      "step": 34000
    },
    {
      "epoch": 1.39,
      "eval_accuracy": 0.7923694779116466,
      "eval_loss": 0.5383234024047852,
      "eval_runtime": 3.1113,
      "eval_samples_per_second": 800.298,
      "eval_steps_per_second": 50.139,
      "step": 34000
    },
    {
      "epoch": 1.41,
      "learning_rate": 3.243919084093872e-06,
      "loss": 0.49,
      "step": 34500
    },
    {
      "epoch": 1.43,
      "learning_rate": 3.218454612125163e-06,
      "loss": 0.5014,
      "step": 35000
    },
    {
      "epoch": 1.45,
      "learning_rate": 3.1930410691003916e-06,
      "loss": 0.4877,
      "step": 35500
    },
    {
      "epoch": 1.47,
      "learning_rate": 3.167576597131682e-06,
      "loss": 0.4817,
      "step": 36000
    },
    {
      "epoch": 1.47,
      "eval_accuracy": 0.7883534136546185,
      "eval_loss": 0.5456702709197998,
      "eval_runtime": 3.1157,
      "eval_samples_per_second": 799.189,
      "eval_steps_per_second": 50.07,
      "step": 36000
    },
    {
      "epoch": 1.49,
      "learning_rate": 3.1421121251629727e-06,
      "loss": 0.4889,
      "step": 36500
    },
    {
      "epoch": 1.51,
      "learning_rate": 3.1166476531942635e-06,
      "loss": 0.4928,
      "step": 37000
    },
    {
      "epoch": 1.53,
      "learning_rate": 3.0911831812255543e-06,
      "loss": 0.492,
      "step": 37500
    },
    {
      "epoch": 1.55,
      "learning_rate": 3.065718709256845e-06,
      "loss": 0.4958,
      "step": 38000
    },
    {
      "epoch": 1.55,
      "eval_accuracy": 0.7983935742971887,
      "eval_loss": 0.5244956612586975,
      "eval_runtime": 3.1141,
      "eval_samples_per_second": 799.583,
      "eval_steps_per_second": 50.094,
      "step": 38000
    },
    {
      "epoch": 1.57,
      "learning_rate": 3.040254237288136e-06,
      "loss": 0.494,
      "step": 38500
    },
    {
      "epoch": 1.59,
      "learning_rate": 3.0147897653194262e-06,
      "loss": 0.4876,
      "step": 39000
    },
    {
      "epoch": 1.61,
      "learning_rate": 2.989376222294655e-06,
      "loss": 0.4983,
      "step": 39500
    },
    {
      "epoch": 1.63,
      "learning_rate": 2.9639626792698826e-06,
      "loss": 0.4993,
      "step": 40000
    },
    {
      "epoch": 1.63,
      "eval_accuracy": 0.7931726907630522,
      "eval_loss": 0.5367791652679443,
      "eval_runtime": 3.1071,
      "eval_samples_per_second": 801.396,
      "eval_steps_per_second": 50.208,
      "step": 40000
    },
    {
      "epoch": 1.65,
      "learning_rate": 2.9384982073011734e-06,
      "loss": 0.4879,
      "step": 40500
    },
    {
      "epoch": 1.67,
      "learning_rate": 2.9130337353324646e-06,
      "loss": 0.4942,
      "step": 41000
    },
    {
      "epoch": 1.69,
      "learning_rate": 2.8875692633637554e-06,
      "loss": 0.4895,
      "step": 41500
    },
    {
      "epoch": 1.71,
      "learning_rate": 2.862104791395046e-06,
      "loss": 0.4861,
      "step": 42000
    },
    {
      "epoch": 1.71,
      "eval_accuracy": 0.7967871485943775,
      "eval_loss": 0.5284856557846069,
      "eval_runtime": 3.0957,
      "eval_samples_per_second": 804.354,
      "eval_steps_per_second": 50.393,
      "step": 42000
    },
    {
      "epoch": 1.73,
      "learning_rate": 2.8366403194263366e-06,
      "loss": 0.4936,
      "step": 42500
    },
    {
      "epoch": 1.75,
      "learning_rate": 2.8111758474576274e-06,
      "loss": 0.4782,
      "step": 43000
    },
    {
      "epoch": 1.77,
      "learning_rate": 2.785711375488918e-06,
      "loss": 0.4857,
      "step": 43500
    },
    {
      "epoch": 1.79,
      "learning_rate": 2.760297832464146e-06,
      "loss": 0.4865,
      "step": 44000
    },
    {
      "epoch": 1.79,
      "eval_accuracy": 0.7967871485943775,
      "eval_loss": 0.5284005403518677,
      "eval_runtime": 3.1113,
      "eval_samples_per_second": 800.304,
      "eval_steps_per_second": 50.14,
      "step": 44000
    },
    {
      "epoch": 1.81,
      "learning_rate": 2.7348333604954367e-06,
      "loss": 0.4913,
      "step": 44500
    },
    {
      "epoch": 1.83,
      "learning_rate": 2.709368888526728e-06,
      "loss": 0.4899,
      "step": 45000
    },
    {
      "epoch": 1.85,
      "learning_rate": 2.6839044165580187e-06,
      "loss": 0.4872,
      "step": 45500
    },
    {
      "epoch": 1.87,
      "learning_rate": 2.6584399445893095e-06,
      "loss": 0.4827,
      "step": 46000
    },
    {
      "epoch": 1.87,
      "eval_accuracy": 0.7927710843373494,
      "eval_loss": 0.5268122553825378,
      "eval_runtime": 3.1094,
      "eval_samples_per_second": 800.807,
      "eval_steps_per_second": 50.171,
      "step": 46000
    },
    {
      "epoch": 1.89,
      "learning_rate": 2.6330264015645373e-06,
      "loss": 0.4844,
      "step": 46500
    },
    {
      "epoch": 1.91,
      "learning_rate": 2.607561929595828e-06,
      "loss": 0.4803,
      "step": 47000
    },
    {
      "epoch": 1.94,
      "learning_rate": 2.582097457627119e-06,
      "loss": 0.4863,
      "step": 47500
    },
    {
      "epoch": 1.96,
      "learning_rate": 2.556632985658409e-06,
      "loss": 0.4706,
      "step": 48000
    },
    {
      "epoch": 1.96,
      "eval_accuracy": 0.7963855421686747,
      "eval_loss": 0.5178616046905518,
      "eval_runtime": 3.1089,
      "eval_samples_per_second": 800.919,
      "eval_steps_per_second": 50.178,
      "step": 48000
    },
    {
      "epoch": 1.98,
      "learning_rate": 2.5311685136897e-06,
      "loss": 0.4921,
      "step": 48500
    },
    {
      "epoch": 2.0,
      "learning_rate": 2.5057549706649286e-06,
      "loss": 0.4877,
      "step": 49000
    },
    {
      "epoch": 2.02,
      "learning_rate": 2.4802904986962194e-06,
      "loss": 0.4357,
      "step": 49500
    },
    {
      "epoch": 2.04,
      "learning_rate": 2.45482602672751e-06,
      "loss": 0.4374,
      "step": 50000
    },
    {
      "epoch": 2.04,
      "eval_accuracy": 0.7995983935742972,
      "eval_loss": 0.5278629064559937,
      "eval_runtime": 3.1508,
      "eval_samples_per_second": 790.269,
      "eval_steps_per_second": 49.511,
      "step": 50000
    },
    {
      "epoch": 2.06,
      "learning_rate": 2.4293615547588006e-06,
      "loss": 0.418,
      "step": 50500
    },
    {
      "epoch": 2.08,
      "learning_rate": 2.4038970827900913e-06,
      "loss": 0.4102,
      "step": 51000
    },
    {
      "epoch": 2.1,
      "learning_rate": 2.3784835397653195e-06,
      "loss": 0.4268,
      "step": 51500
    },
    {
      "epoch": 2.12,
      "learning_rate": 2.3530190677966103e-06,
      "loss": 0.4238,
      "step": 52000
    },
    {
      "epoch": 2.12,
      "eval_accuracy": 0.7911646586345381,
      "eval_loss": 0.5376625061035156,
      "eval_runtime": 3.1036,
      "eval_samples_per_second": 802.307,
      "eval_steps_per_second": 50.265,
      "step": 52000
    },
    {
      "epoch": 2.14,
      "learning_rate": 2.327554595827901e-06,
      "loss": 0.419,
      "step": 52500
    },
    {
      "epoch": 2.16,
      "learning_rate": 2.302090123859192e-06,
      "loss": 0.4006,
      "step": 53000
    },
    {
      "epoch": 2.18,
      "learning_rate": 2.2766765808344197e-06,
      "loss": 0.4142,
      "step": 53500
    },
    {
      "epoch": 2.2,
      "learning_rate": 2.251212108865711e-06,
      "loss": 0.4128,
      "step": 54000
    },
    {
      "epoch": 2.2,
      "eval_accuracy": 0.7971887550200804,
      "eval_loss": 0.5341054797172546,
      "eval_runtime": 3.1779,
      "eval_samples_per_second": 783.533,
      "eval_steps_per_second": 49.089,
      "step": 54000
    },
    {
      "epoch": 2.22,
      "learning_rate": 2.2257476368970017e-06,
      "loss": 0.4169,
      "step": 54500
    },
    {
      "epoch": 2.24,
      "learning_rate": 2.200283164928292e-06,
      "loss": 0.4219,
      "step": 55000
    },
    {
      "epoch": 2.26,
      "learning_rate": 2.174818692959583e-06,
      "loss": 0.3998,
      "step": 55500
    },
    {
      "epoch": 2.28,
      "learning_rate": 2.149405149934811e-06,
      "loss": 0.4079,
      "step": 56000
    },
    {
      "epoch": 2.28,
      "eval_accuracy": 0.8036144578313253,
      "eval_loss": 0.53012615442276,
      "eval_runtime": 3.1161,
      "eval_samples_per_second": 799.075,
      "eval_steps_per_second": 50.063,
      "step": 56000
    },
    {
      "epoch": 2.3,
      "learning_rate": 2.123940677966102e-06,
      "loss": 0.4139,
      "step": 56500
    },
    {
      "epoch": 2.32,
      "learning_rate": 2.0984762059973926e-06,
      "loss": 0.4117,
      "step": 57000
    },
    {
      "epoch": 2.34,
      "learning_rate": 2.0730117340286834e-06,
      "loss": 0.4144,
      "step": 57500
    },
    {
      "epoch": 2.36,
      "learning_rate": 2.0475981910039116e-06,
      "loss": 0.4219,
      "step": 58000
    },
    {
      "epoch": 2.36,
      "eval_accuracy": 0.8,
      "eval_loss": 0.5489892363548279,
      "eval_runtime": 3.1097,
      "eval_samples_per_second": 800.724,
      "eval_steps_per_second": 50.166,
      "step": 58000
    },
    {
      "epoch": 2.38,
      "learning_rate": 2.0221337190352024e-06,
      "loss": 0.4231,
      "step": 58500
    },
    {
      "epoch": 2.4,
      "learning_rate": 1.996669247066493e-06,
      "loss": 0.4054,
      "step": 59000
    },
    {
      "epoch": 2.42,
      "learning_rate": 1.9712047750977835e-06,
      "loss": 0.4166,
      "step": 59500
    },
    {
      "epoch": 2.44,
      "learning_rate": 1.9457912320730117e-06,
      "loss": 0.3995,
      "step": 60000
    },
    {
      "epoch": 2.44,
      "eval_accuracy": 0.8028112449799196,
      "eval_loss": 0.5354592204093933,
      "eval_runtime": 3.1379,
      "eval_samples_per_second": 793.526,
      "eval_steps_per_second": 49.715,
      "step": 60000
    },
    {
      "epoch": 2.46,
      "learning_rate": 1.9203267601043025e-06,
      "loss": 0.4244,
      "step": 60500
    },
    {
      "epoch": 2.49,
      "learning_rate": 1.8948622881355933e-06,
      "loss": 0.412,
      "step": 61000
    },
    {
      "epoch": 2.51,
      "learning_rate": 1.869397816166884e-06,
      "loss": 0.4108,
      "step": 61500
    },
    {
      "epoch": 2.53,
      "learning_rate": 1.8439333441981749e-06,
      "loss": 0.4181,
      "step": 62000
    },
    {
      "epoch": 2.53,
      "eval_accuracy": 0.802008032128514,
      "eval_loss": 0.5300112366676331,
      "eval_runtime": 3.1096,
      "eval_samples_per_second": 800.752,
      "eval_steps_per_second": 50.168,
      "step": 62000
    },
    {
      "epoch": 2.55,
      "learning_rate": 1.8184688722294654e-06,
      "loss": 0.4081,
      "step": 62500
    },
    {
      "epoch": 2.57,
      "learning_rate": 1.7930044002607562e-06,
      "loss": 0.4048,
      "step": 63000
    },
    {
      "epoch": 2.59,
      "learning_rate": 1.7675399282920472e-06,
      "loss": 0.429,
      "step": 63500
    },
    {
      "epoch": 2.61,
      "learning_rate": 1.7420754563233378e-06,
      "loss": 0.4179,
      "step": 64000
    },
    {
      "epoch": 2.61,
      "eval_accuracy": 0.7963855421686747,
      "eval_loss": 0.5444837212562561,
      "eval_runtime": 3.1167,
      "eval_samples_per_second": 798.922,
      "eval_steps_per_second": 50.053,
      "step": 64000
    },
    {
      "epoch": 2.63,
      "learning_rate": 1.7166109843546286e-06,
      "loss": 0.4114,
      "step": 64500
    },
    {
      "epoch": 2.65,
      "learning_rate": 1.6911465123859194e-06,
      "loss": 0.407,
      "step": 65000
    },
    {
      "epoch": 2.67,
      "learning_rate": 1.66568204041721e-06,
      "loss": 0.4128,
      "step": 65500
    },
    {
      "epoch": 2.69,
      "learning_rate": 1.6402684973924382e-06,
      "loss": 0.4085,
      "step": 66000
    },
    {
      "epoch": 2.69,
      "eval_accuracy": 0.795582329317269,
      "eval_loss": 0.5385735034942627,
      "eval_runtime": 3.1063,
      "eval_samples_per_second": 801.601,
      "eval_steps_per_second": 50.221,
      "step": 66000
    },
    {
      "epoch": 2.71,
      "learning_rate": 1.6148040254237292e-06,
      "loss": 0.4036,
      "step": 66500
    },
    {
      "epoch": 2.73,
      "learning_rate": 1.5893395534550195e-06,
      "loss": 0.4169,
      "step": 67000
    },
    {
      "epoch": 2.75,
      "learning_rate": 1.5638750814863105e-06,
      "loss": 0.4301,
      "step": 67500
    },
    {
      "epoch": 2.77,
      "learning_rate": 1.5384106095176013e-06,
      "loss": 0.4183,
      "step": 68000
    },
    {
      "epoch": 2.77,
      "eval_accuracy": 0.8004016064257028,
      "eval_loss": 0.5275784730911255,
      "eval_runtime": 3.1078,
      "eval_samples_per_second": 801.199,
      "eval_steps_per_second": 50.196,
      "step": 68000
    },
    {
      "epoch": 2.79,
      "learning_rate": 1.5129970664928293e-06,
      "loss": 0.4114,
      "step": 68500
    },
    {
      "epoch": 2.81,
      "learning_rate": 1.48753259452412e-06,
      "loss": 0.4236,
      "step": 69000
    },
    {
      "epoch": 2.83,
      "learning_rate": 1.4620681225554109e-06,
      "loss": 0.4179,
      "step": 69500
    },
    {
      "epoch": 2.85,
      "learning_rate": 1.4366036505867014e-06,
      "loss": 0.4132,
      "step": 70000
    },
    {
      "epoch": 2.85,
      "eval_accuracy": 0.7963855421686747,
      "eval_loss": 0.5376485586166382,
      "eval_runtime": 3.1089,
      "eval_samples_per_second": 800.921,
      "eval_steps_per_second": 50.178,
      "step": 70000
    },
    {
      "epoch": 2.87,
      "learning_rate": 1.4111901075619296e-06,
      "loss": 0.4183,
      "step": 70500
    },
    {
      "epoch": 2.89,
      "learning_rate": 1.3857765645371578e-06,
      "loss": 0.409,
      "step": 71000
    },
    {
      "epoch": 2.91,
      "learning_rate": 1.3603120925684484e-06,
      "loss": 0.4113,
      "step": 71500
    },
    {
      "epoch": 2.93,
      "learning_rate": 1.3348476205997394e-06,
      "loss": 0.4031,
      "step": 72000
    },
    {
      "epoch": 2.93,
      "eval_accuracy": 0.7987951807228916,
      "eval_loss": 0.5440085530281067,
      "eval_runtime": 3.1049,
      "eval_samples_per_second": 801.949,
      "eval_steps_per_second": 50.243,
      "step": 72000
    },
    {
      "epoch": 2.95,
      "learning_rate": 1.3093831486310302e-06,
      "loss": 0.4153,
      "step": 72500
    },
    {
      "epoch": 2.97,
      "learning_rate": 1.2839186766623208e-06,
      "loss": 0.4092,
      "step": 73000
    },
    {
      "epoch": 2.99,
      "learning_rate": 1.2584542046936116e-06,
      "loss": 0.4223,
      "step": 73500
    },
    {
      "epoch": 3.01,
      "learning_rate": 1.2329897327249023e-06,
      "loss": 0.3561,
      "step": 74000
    },
    {
      "epoch": 3.01,
      "eval_accuracy": 0.8032128514056225,
      "eval_loss": 0.5515400171279907,
      "eval_runtime": 3.1138,
      "eval_samples_per_second": 799.666,
      "eval_steps_per_second": 50.1,
      "step": 74000
    },
    {
      "epoch": 3.04,
      "learning_rate": 1.2075252607561931e-06,
      "loss": 0.3444,
      "step": 74500
    },
    {
      "epoch": 3.06,
      "learning_rate": 1.1821117177314211e-06,
      "loss": 0.3447,
      "step": 75000
    },
    {
      "epoch": 3.08,
      "learning_rate": 1.156647245762712e-06,
      "loss": 0.3572,
      "step": 75500
    },
    {
      "epoch": 3.1,
      "learning_rate": 1.1311827737940027e-06,
      "loss": 0.3453,
      "step": 76000
    },
    {
      "epoch": 3.1,
      "eval_accuracy": 0.8036144578313253,
      "eval_loss": 0.562953770160675,
      "eval_runtime": 3.1189,
      "eval_samples_per_second": 798.371,
      "eval_steps_per_second": 50.018,
      "step": 76000
    },
    {
      "epoch": 3.12,
      "learning_rate": 1.1057183018252935e-06,
      "loss": 0.353,
      "step": 76500
    },
    {
      "epoch": 3.14,
      "learning_rate": 1.0803047588005217e-06,
      "loss": 0.3489,
      "step": 77000
    },
    {
      "epoch": 3.16,
      "learning_rate": 1.0548402868318123e-06,
      "loss": 0.3487,
      "step": 77500
    },
    {
      "epoch": 3.18,
      "learning_rate": 1.029375814863103e-06,
      "loss": 0.3625,
      "step": 78000
    },
    {
      "epoch": 3.18,
      "eval_accuracy": 0.8012048192771084,
      "eval_loss": 0.5594542026519775,
      "eval_runtime": 3.1009,
      "eval_samples_per_second": 802.99,
      "eval_steps_per_second": 50.308,
      "step": 78000
    },
    {
      "epoch": 3.2,
      "learning_rate": 1.0039113428943938e-06,
      "loss": 0.3522,
      "step": 78500
    },
    {
      "epoch": 3.22,
      "learning_rate": 9.784468709256846e-07,
      "loss": 0.3526,
      "step": 79000
    },
    {
      "epoch": 3.24,
      "learning_rate": 9.530333279009127e-07,
      "loss": 0.3463,
      "step": 79500
    },
    {
      "epoch": 3.26,
      "learning_rate": 9.275688559322035e-07,
      "loss": 0.3483,
      "step": 80000
    },
    {
      "epoch": 3.26,
      "eval_accuracy": 0.8024096385542169,
      "eval_loss": 0.5727831125259399,
      "eval_runtime": 3.1106,
      "eval_samples_per_second": 800.501,
      "eval_steps_per_second": 50.152,
      "step": 80000
    },
    {
      "epoch": 3.28,
      "learning_rate": 9.021043839634942e-07,
      "loss": 0.3614,
      "step": 80500
    },
    {
      "epoch": 3.3,
      "learning_rate": 8.76639911994785e-07,
      "loss": 0.3464,
      "step": 81000
    },
    {
      "epoch": 3.32,
      "learning_rate": 8.511754400260756e-07,
      "loss": 0.3534,
      "step": 81500
    },
    {
      "epoch": 3.34,
      "learning_rate": 8.257618970013039e-07,
      "loss": 0.3452,
      "step": 82000
    },
    {
      "epoch": 3.34,
      "eval_accuracy": 0.8032128514056225,
      "eval_loss": 0.5750517845153809,
      "eval_runtime": 3.0909,
      "eval_samples_per_second": 805.579,
      "eval_steps_per_second": 50.47,
      "step": 82000
    },
    {
      "epoch": 3.36,
      "learning_rate": 8.002974250325945e-07,
      "loss": 0.3399,
      "step": 82500
    },
    {
      "epoch": 3.38,
      "learning_rate": 7.748329530638854e-07,
      "loss": 0.3474,
      "step": 83000
    },
    {
      "epoch": 3.4,
      "learning_rate": 7.493684810951761e-07,
      "loss": 0.3629,
      "step": 83500
    },
    {
      "epoch": 3.42,
      "learning_rate": 7.239040091264668e-07,
      "loss": 0.3481,
      "step": 84000
    },
    {
      "epoch": 3.42,
      "eval_accuracy": 0.7975903614457831,
      "eval_loss": 0.5842192769050598,
      "eval_runtime": 3.1136,
      "eval_samples_per_second": 799.707,
      "eval_steps_per_second": 50.102,
      "step": 84000
    },
    {
      "epoch": 3.44,
      "learning_rate": 6.984395371577575e-07,
      "loss": 0.347,
      "step": 84500
    },
    {
      "epoch": 3.46,
      "learning_rate": 6.729750651890484e-07,
      "loss": 0.3659,
      "step": 85000
    },
    {
      "epoch": 3.48,
      "learning_rate": 6.47510593220339e-07,
      "loss": 0.3414,
      "step": 85500
    },
    {
      "epoch": 3.5,
      "learning_rate": 6.220970501955672e-07,
      "loss": 0.3526,
      "step": 86000
    },
    {
      "epoch": 3.5,
      "eval_accuracy": 0.7983935742971887,
      "eval_loss": 0.580718994140625,
      "eval_runtime": 3.0922,
      "eval_samples_per_second": 805.245,
      "eval_steps_per_second": 50.449,
      "step": 86000
    },
    {
      "epoch": 3.52,
      "learning_rate": 5.966325782268579e-07,
      "loss": 0.3502,
      "step": 86500
    },
    {
      "epoch": 3.54,
      "learning_rate": 5.712190352020861e-07,
      "loss": 0.3522,
      "step": 87000
    },
    {
      "epoch": 3.57,
      "learning_rate": 5.457545632333768e-07,
      "loss": 0.3451,
      "step": 87500
    },
    {
      "epoch": 3.59,
      "learning_rate": 5.202900912646676e-07,
      "loss": 0.3513,
      "step": 88000
    },
    {
      "epoch": 3.59,
      "eval_accuracy": 0.8016064257028113,
      "eval_loss": 0.5780753493309021,
      "eval_runtime": 3.1077,
      "eval_samples_per_second": 801.223,
      "eval_steps_per_second": 50.197,
      "step": 88000
    },
    {
      "epoch": 3.61,
      "learning_rate": 4.948256192959583e-07,
      "loss": 0.3386,
      "step": 88500
    },
    {
      "epoch": 3.63,
      "learning_rate": 4.6936114732724906e-07,
      "loss": 0.3465,
      "step": 89000
    },
    {
      "epoch": 3.65,
      "learning_rate": 4.4389667535853984e-07,
      "loss": 0.3474,
      "step": 89500
    },
    {
      "epoch": 3.67,
      "learning_rate": 4.184322033898305e-07,
      "loss": 0.3288,
      "step": 90000
    },
    {
      "epoch": 3.67,
      "eval_accuracy": 0.8008032128514057,
      "eval_loss": 0.5857793688774109,
      "eval_runtime": 3.11,
      "eval_samples_per_second": 800.647,
      "eval_steps_per_second": 50.161,
      "step": 90000
    },
    {
      "epoch": 3.69,
      "learning_rate": 3.9306958930899614e-07,
      "loss": 0.3444,
      "step": 90500
    },
    {
      "epoch": 3.71,
      "learning_rate": 3.676051173402868e-07,
      "loss": 0.3373,
      "step": 91000
    },
    {
      "epoch": 3.73,
      "learning_rate": 3.421406453715776e-07,
      "loss": 0.3512,
      "step": 91500
    },
    {
      "epoch": 3.75,
      "learning_rate": 3.166761734028683e-07,
      "loss": 0.3599,
      "step": 92000
    },
    {
      "epoch": 3.75,
      "eval_accuracy": 0.8016064257028113,
      "eval_loss": 0.5658962726593018,
      "eval_runtime": 3.112,
      "eval_samples_per_second": 800.117,
      "eval_steps_per_second": 50.128,
      "step": 92000
    },
    {
      "epoch": 3.77,
      "learning_rate": 2.9121170143415907e-07,
      "loss": 0.3535,
      "step": 92500
    },
    {
      "epoch": 3.79,
      "learning_rate": 2.657472294654498e-07,
      "loss": 0.3404,
      "step": 93000
    },
    {
      "epoch": 3.81,
      "learning_rate": 2.4028275749674054e-07,
      "loss": 0.3452,
      "step": 93500
    },
    {
      "epoch": 3.83,
      "learning_rate": 2.148182855280313e-07,
      "loss": 0.3419,
      "step": 94000
    },
    {
      "epoch": 3.83,
      "eval_accuracy": 0.8032128514056225,
      "eval_loss": 0.575123131275177,
      "eval_runtime": 3.1066,
      "eval_samples_per_second": 801.524,
      "eval_steps_per_second": 50.216,
      "step": 94000
    },
    {
      "epoch": 3.85,
      "learning_rate": 1.8935381355932206e-07,
      "loss": 0.3518,
      "step": 94500
    },
    {
      "epoch": 3.87,
      "learning_rate": 1.638893415906128e-07,
      "loss": 0.3444,
      "step": 95000
    },
    {
      "epoch": 3.89,
      "learning_rate": 1.3842486962190353e-07,
      "loss": 0.3422,
      "step": 95500
    },
    {
      "epoch": 3.91,
      "learning_rate": 1.1296039765319427e-07,
      "loss": 0.3502,
      "step": 96000
    },
    {
      "epoch": 3.91,
      "eval_accuracy": 0.8036144578313253,
      "eval_loss": 0.5685034394264221,
      "eval_runtime": 3.1416,
      "eval_samples_per_second": 792.595,
      "eval_steps_per_second": 49.657,
      "step": 96000
    },
    {
      "epoch": 3.93,
      "learning_rate": 8.754685462842242e-08,
      "loss": 0.3373,
      "step": 96500
    },
    {
      "epoch": 3.95,
      "learning_rate": 6.208238265971317e-08,
      "loss": 0.3587,
      "step": 97000
    },
    {
      "epoch": 3.97,
      "learning_rate": 3.6617910691003915e-08,
      "loss": 0.3398,
      "step": 97500
    },
    {
      "epoch": 3.99,
      "learning_rate": 1.1153438722294656e-08,
      "loss": 0.3443,
      "step": 98000
    },
    {
      "epoch": 3.99,
      "eval_accuracy": 0.8016064257028113,
      "eval_loss": 0.5707261562347412,
      "eval_runtime": 3.1118,
      "eval_samples_per_second": 800.188,
      "eval_steps_per_second": 50.132,
      "step": 98000
    },
    {
      "epoch": 4.0,
      "step": 98176,
      "total_flos": 1.1341991182001568e+16,
      "train_loss": 0.4772139054466879,
      "train_runtime": 9899.244,
      "train_samples_per_second": 158.68,
      "train_steps_per_second": 9.918
    }
  ],
  "max_steps": 98176,
  "num_train_epochs": 4,
  "total_flos": 1.1341991182001568e+16,
  "trial_name": null,
  "trial_params": null
}