{
  "best_metric": 1.8623046875,
  "best_model_checkpoint": "bert_12_layer_model_v2_complete_training/checkpoint-450000",
  "epoch": 5.0,
  "global_step": 457720,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 3.3715366784496464e-05,
      "loss": 7.3168,
      "step": 500
    },
    {
      "epoch": 0.01,
      "learning_rate": 3.748913176609214e-05,
      "loss": 6.6227,
      "step": 1000
    },
    {
      "epoch": 0.02,
      "learning_rate": 3.9686644919105525e-05,
      "loss": 6.5293,
      "step": 1500
    },
    {
      "epoch": 0.02,
      "learning_rate": 4.12520067118919e-05,
      "loss": 6.4621,
      "step": 2000
    },
    {
      "epoch": 0.03,
      "learning_rate": 4.246120561428155e-05,
      "loss": 6.423,
      "step": 2500
    },
    {
      "epoch": 0.03,
      "learning_rate": 4.345314745008792e-05,
      "loss": 6.3848,
      "step": 3000
    },
    {
      "epoch": 0.04,
      "learning_rate": 4.428842793791915e-05,
      "loss": 6.3631,
      "step": 3500
    },
    {
      "epoch": 0.04,
      "learning_rate": 4.501488165769166e-05,
      "loss": 6.3323,
      "step": 4000
    },
    {
      "epoch": 0.05,
      "learning_rate": 4.5653079262541536e-05,
      "loss": 6.3111,
      "step": 4500
    },
    {
      "epoch": 0.05,
      "learning_rate": 4.622625682029237e-05,
      "loss": 6.3021,
      "step": 5000
    },
    {
      "epoch": 0.06,
      "learning_rate": 4.6742676283733686e-05,
      "loss": 6.277,
      "step": 5500
    },
    {
      "epoch": 0.07,
      "learning_rate": 4.7216022395887685e-05,
      "loss": 6.2674,
      "step": 6000
    },
    {
      "epoch": 0.07,
      "learning_rate": 4.7649711804211924e-05,
      "loss": 6.2475,
      "step": 6500
    },
    {
      "epoch": 0.08,
      "learning_rate": 4.805285726627034e-05,
      "loss": 6.2337,
      "step": 7000
    },
    {
      "epoch": 0.08,
      "learning_rate": 4.842667223534965e-05,
      "loss": 6.2266,
      "step": 7500
    },
    {
      "epoch": 0.09,
      "learning_rate": 4.8777756603491424e-05,
      "loss": 6.2155,
      "step": 8000
    },
    {
      "epoch": 0.09,
      "learning_rate": 4.910622835154613e-05,
      "loss": 6.2127,
      "step": 8500
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.941716313408369e-05,
      "loss": 6.2033,
      "step": 9000
    },
    {
      "epoch": 0.1,
      "learning_rate": 4.971010421672582e-05,
      "loss": 6.1884,
      "step": 9500
    },
    {
      "epoch": 0.11,
      "learning_rate": 4.9989131766092136e-05,
      "loss": 6.1798,
      "step": 10000
    },
    {
      "epoch": 0.11,
      "eval_accuracy": 0.14852960767779255,
      "eval_loss": 6.171875,
      "eval_runtime": 1078.8913,
      "eval_samples_per_second": 285.808,
      "eval_steps_per_second": 4.467,
      "step": 10000
    },
    {
      "epoch": 0.11,
      "learning_rate": 4.994673009916912e-05,
      "loss": 6.1762,
      "step": 10500
    },
    {
      "epoch": 0.12,
      "learning_rate": 4.9890891628696506e-05,
      "loss": 6.1649,
      "step": 11000
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.983527651210578e-05,
      "loss": 6.1515,
      "step": 11500
    },
    {
      "epoch": 0.13,
      "learning_rate": 4.9779438041633164e-05,
      "loss": 6.1443,
      "step": 12000
    },
    {
      "epoch": 0.14,
      "learning_rate": 4.972382292504244e-05,
      "loss": 6.142,
      "step": 12500
    },
    {
      "epoch": 0.14,
      "learning_rate": 4.966798445456982e-05,
      "loss": 6.1324,
      "step": 13000
    },
    {
      "epoch": 0.15,
      "learning_rate": 4.96123693379791e-05,
      "loss": 6.1284,
      "step": 13500
    },
    {
      "epoch": 0.15,
      "learning_rate": 4.955653086750648e-05,
      "loss": 6.1214,
      "step": 14000
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.9500915750915755e-05,
      "loss": 6.1157,
      "step": 14500
    },
    {
      "epoch": 0.16,
      "learning_rate": 4.944507728044314e-05,
      "loss": 6.1087,
      "step": 15000
    },
    {
      "epoch": 0.17,
      "learning_rate": 4.938946216385241e-05,
      "loss": 6.0982,
      "step": 15500
    },
    {
      "epoch": 0.17,
      "learning_rate": 4.9333623693379796e-05,
      "loss": 6.0918,
      "step": 16000
    },
    {
      "epoch": 0.18,
      "learning_rate": 4.9278008576789064e-05,
      "loss": 6.0853,
      "step": 16500
    },
    {
      "epoch": 0.19,
      "learning_rate": 4.9222170106316454e-05,
      "loss": 6.0825,
      "step": 17000
    },
    {
      "epoch": 0.19,
      "learning_rate": 4.916655498972572e-05,
      "loss": 6.075,
      "step": 17500
    },
    {
      "epoch": 0.2,
      "learning_rate": 4.9110716519253105e-05,
      "loss": 6.0639,
      "step": 18000
    },
    {
      "epoch": 0.2,
      "learning_rate": 4.905510140266238e-05,
      "loss": 6.0632,
      "step": 18500
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.8999262932189763e-05,
      "loss": 6.0566,
      "step": 19000
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.894364781559904e-05,
      "loss": 6.0534,
      "step": 19500
    },
    {
      "epoch": 0.22,
      "learning_rate": 4.888780934512642e-05,
      "loss": 6.0527,
      "step": 20000
    },
    {
      "epoch": 0.22,
      "eval_accuracy": 0.1501731560033556,
      "eval_loss": 6.046875,
      "eval_runtime": 1109.6875,
      "eval_samples_per_second": 277.876,
      "eval_steps_per_second": 4.343,
      "step": 20000
    },
    {
      "epoch": 0.22,
      "learning_rate": 4.8832194228535696e-05,
      "loss": 6.0424,
      "step": 20500
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.877635575806307e-05,
      "loss": 6.0454,
      "step": 21000
    },
    {
      "epoch": 0.23,
      "learning_rate": 4.8720740641472354e-05,
      "loss": 6.038,
      "step": 21500
    },
    {
      "epoch": 0.24,
      "learning_rate": 4.866490217099973e-05,
      "loss": 6.0292,
      "step": 22000
    },
    {
      "epoch": 0.25,
      "learning_rate": 4.8609287054409006e-05,
      "loss": 6.0285,
      "step": 22500
    },
    {
      "epoch": 0.25,
      "learning_rate": 4.855344858393639e-05,
      "loss": 6.0209,
      "step": 23000
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.8497833467345664e-05,
      "loss": 6.0183,
      "step": 23500
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.844199499687305e-05,
      "loss": 6.0132,
      "step": 24000
    },
    {
      "epoch": 0.27,
      "learning_rate": 4.838637988028232e-05,
      "loss": 6.0076,
      "step": 24500
    },
    {
      "epoch": 0.27,
      "learning_rate": 4.8330541409809705e-05,
      "loss": 5.9644,
      "step": 25000
    },
    {
      "epoch": 0.28,
      "learning_rate": 4.827492629321898e-05,
      "loss": 5.8913,
      "step": 25500
    },
    {
      "epoch": 0.28,
      "learning_rate": 4.821908782274636e-05,
      "loss": 5.8423,
      "step": 26000
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.816347270615564e-05,
      "loss": 5.8085,
      "step": 26500
    },
    {
      "epoch": 0.29,
      "learning_rate": 4.810763423568302e-05,
      "loss": 5.7797,
      "step": 27000
    },
    {
      "epoch": 0.3,
      "learning_rate": 4.805201911909229e-05,
      "loss": 5.7515,
      "step": 27500
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.799618064861968e-05,
      "loss": 5.7142,
      "step": 28000
    },
    {
      "epoch": 0.31,
      "learning_rate": 4.794056553202895e-05,
      "loss": 5.6929,
      "step": 28500
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.788472706155634e-05,
      "loss": 5.6603,
      "step": 29000
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.7829111944965605e-05,
      "loss": 5.6395,
      "step": 29500
    },
    {
      "epoch": 0.33,
      "learning_rate": 4.777327347449299e-05,
      "loss": 5.6176,
      "step": 30000
    },
    {
      "epoch": 0.33,
      "eval_accuracy": 0.1772248767905883,
      "eval_loss": 5.5703125,
      "eval_runtime": 1165.3244,
      "eval_samples_per_second": 264.61,
      "eval_steps_per_second": 4.135,
      "step": 30000
    },
    {
      "epoch": 0.33,
      "learning_rate": 4.771765835790226e-05,
      "loss": 5.5758,
      "step": 30500
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.7661819887429646e-05,
      "loss": 5.4849,
      "step": 31000
    },
    {
      "epoch": 0.34,
      "learning_rate": 4.760620477083892e-05,
      "loss": 5.348,
      "step": 31500
    },
    {
      "epoch": 0.35,
      "learning_rate": 4.75503663003663e-05,
      "loss": 5.2098,
      "step": 32000
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.749475118377558e-05,
      "loss": 5.1005,
      "step": 32500
    },
    {
      "epoch": 0.36,
      "learning_rate": 4.7438912713302956e-05,
      "loss": 5.0023,
      "step": 33000
    },
    {
      "epoch": 0.37,
      "learning_rate": 4.738329759671224e-05,
      "loss": 4.9191,
      "step": 33500
    },
    {
      "epoch": 0.37,
      "learning_rate": 4.7327459126239614e-05,
      "loss": 4.8237,
      "step": 34000
    },
    {
      "epoch": 0.38,
      "learning_rate": 4.727184400964889e-05,
      "loss": 4.7396,
      "step": 34500
    },
    {
      "epoch": 0.38,
      "learning_rate": 4.721600553917627e-05,
      "loss": 4.6683,
      "step": 35000
    },
    {
      "epoch": 0.39,
      "learning_rate": 4.7160390422585546e-05,
      "loss": 4.5571,
      "step": 35500
    },
    {
      "epoch": 0.39,
      "learning_rate": 4.710455195211293e-05,
      "loss": 4.419,
      "step": 36000
    },
    {
      "epoch": 0.4,
      "learning_rate": 4.7048936835522204e-05,
      "loss": 4.2976,
      "step": 36500
    },
    {
      "epoch": 0.4,
      "learning_rate": 4.699309836504959e-05,
      "loss": 4.2061,
      "step": 37000
    },
    {
      "epoch": 0.41,
      "learning_rate": 4.6937483248458856e-05,
      "loss": 4.1266,
      "step": 37500
    },
    {
      "epoch": 0.42,
      "learning_rate": 4.6881644777986246e-05,
      "loss": 4.0604,
      "step": 38000
    },
    {
      "epoch": 0.42,
      "learning_rate": 4.6826029661395514e-05,
      "loss": 4.0129,
      "step": 38500
    },
    {
      "epoch": 0.43,
      "learning_rate": 4.6770191190922904e-05,
      "loss": 3.96,
      "step": 39000
    },
    {
      "epoch": 0.43,
      "learning_rate": 4.671457607433217e-05,
      "loss": 3.9183,
      "step": 39500
    },
    {
      "epoch": 0.44,
      "learning_rate": 4.665873760385956e-05,
      "loss": 3.8786,
      "step": 40000
    },
    {
      "epoch": 0.44,
      "eval_accuracy": 0.3851306885682685,
      "eval_loss": 3.744140625,
      "eval_runtime": 1141.9652,
      "eval_samples_per_second": 270.022,
      "eval_steps_per_second": 4.22,
      "step": 40000
    },
    {
      "epoch": 0.44,
      "learning_rate": 4.660312248726883e-05,
      "loss": 3.8396,
      "step": 40500
    },
    {
      "epoch": 0.45,
      "learning_rate": 4.654728401679622e-05,
      "loss": 3.8043,
      "step": 41000
    },
    {
      "epoch": 0.45,
      "learning_rate": 4.649166890020549e-05,
      "loss": 3.7731,
      "step": 41500
    },
    {
      "epoch": 0.46,
      "learning_rate": 4.643583042973287e-05,
      "loss": 3.7382,
      "step": 42000
    },
    {
      "epoch": 0.46,
      "learning_rate": 4.6380215313142146e-05,
      "loss": 3.7107,
      "step": 42500
    },
    {
      "epoch": 0.47,
      "learning_rate": 4.632437684266953e-05,
      "loss": 3.6783,
      "step": 43000
    },
    {
      "epoch": 0.48,
      "learning_rate": 4.6268761726078804e-05,
      "loss": 3.6592,
      "step": 43500
    },
    {
      "epoch": 0.48,
      "learning_rate": 4.621292325560618e-05,
      "loss": 3.6362,
      "step": 44000
    },
    {
      "epoch": 0.49,
      "learning_rate": 4.615730813901546e-05,
      "loss": 3.6159,
      "step": 44500
    },
    {
      "epoch": 0.49,
      "learning_rate": 4.610146966854284e-05,
      "loss": 3.5905,
      "step": 45000
    },
    {
      "epoch": 0.5,
      "learning_rate": 4.604585455195212e-05,
      "loss": 3.5683,
      "step": 45500
    },
    {
      "epoch": 0.5,
      "learning_rate": 4.5990016081479496e-05,
      "loss": 3.5432,
      "step": 46000
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.593440096488877e-05,
      "loss": 3.527,
      "step": 46500
    },
    {
      "epoch": 0.51,
      "learning_rate": 4.5878562494416154e-05,
      "loss": 3.5073,
      "step": 47000
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.582294737782543e-05,
      "loss": 3.4948,
      "step": 47500
    },
    {
      "epoch": 0.52,
      "learning_rate": 4.576710890735281e-05,
      "loss": 3.4776,
      "step": 48000
    },
    {
      "epoch": 0.53,
      "learning_rate": 4.571149379076208e-05,
      "loss": 3.4541,
      "step": 48500
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.565565532028947e-05,
      "loss": 3.443,
      "step": 49000
    },
    {
      "epoch": 0.54,
      "learning_rate": 4.560004020369874e-05,
      "loss": 3.4296,
      "step": 49500
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.554420173322613e-05,
      "loss": 3.4104,
      "step": 50000
    },
    {
      "epoch": 0.55,
      "eval_accuracy": 0.4326803997069381,
      "eval_loss": 3.310546875,
      "eval_runtime": 1153.5787,
      "eval_samples_per_second": 267.304,
      "eval_steps_per_second": 4.177,
      "step": 50000
    },
    {
      "epoch": 0.55,
      "learning_rate": 4.5488586616635397e-05,
      "loss": 3.3905,
      "step": 50500
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.5432748146162787e-05,
      "loss": 3.3848,
      "step": 51000
    },
    {
      "epoch": 0.56,
      "learning_rate": 4.5377133029572055e-05,
      "loss": 3.3768,
      "step": 51500
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.5321294559099445e-05,
      "loss": 3.3554,
      "step": 52000
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.526567944250871e-05,
      "loss": 3.3403,
      "step": 52500
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.5209840972036096e-05,
      "loss": 3.3207,
      "step": 53000
    },
    {
      "epoch": 0.58,
      "learning_rate": 4.515422585544537e-05,
      "loss": 3.3162,
      "step": 53500
    },
    {
      "epoch": 0.59,
      "learning_rate": 4.5098387384972754e-05,
      "loss": 3.2997,
      "step": 54000
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.504277226838203e-05,
      "loss": 3.2948,
      "step": 54500
    },
    {
      "epoch": 0.6,
      "learning_rate": 4.498693379790941e-05,
      "loss": 3.2763,
      "step": 55000
    },
    {
      "epoch": 0.61,
      "learning_rate": 4.493131868131869e-05,
      "loss": 3.2682,
      "step": 55500
    },
    {
      "epoch": 0.61,
      "learning_rate": 4.487548021084606e-05,
      "loss": 3.2571,
      "step": 56000
    },
    {
      "epoch": 0.62,
      "learning_rate": 4.4819865094255345e-05,
      "loss": 3.243,
      "step": 56500
    },
    {
      "epoch": 0.62,
      "learning_rate": 4.476402662378272e-05,
      "loss": 3.2327,
      "step": 57000
    },
    {
      "epoch": 0.63,
      "learning_rate": 4.4708411507192e-05,
      "loss": 3.229,
      "step": 57500
    },
    {
      "epoch": 0.63,
      "learning_rate": 4.465257303671938e-05,
      "loss": 3.2023,
      "step": 58000
    },
    {
      "epoch": 0.64,
      "learning_rate": 4.4596957920128654e-05,
      "loss": 3.2021,
      "step": 58500
    },
    {
      "epoch": 0.64,
      "learning_rate": 4.454111944965604e-05,
      "loss": 3.1926,
      "step": 59000
    },
    {
      "epoch": 0.65,
      "learning_rate": 4.448550433306531e-05,
      "loss": 3.1862,
      "step": 59500
    },
    {
      "epoch": 0.66,
      "learning_rate": 4.4429665862592695e-05,
      "loss": 3.1802,
      "step": 60000
    },
    {
      "epoch": 0.66,
      "eval_accuracy": 0.4600636447202696,
      "eval_loss": 3.078125,
      "eval_runtime": 1125.3252,
      "eval_samples_per_second": 274.015,
      "eval_steps_per_second": 4.282,
      "step": 60000
    },
    {
      "epoch": 0.66,
      "learning_rate": 4.437405074600196e-05,
      "loss": 3.164,
      "step": 60500
    },
    {
      "epoch": 0.67,
      "learning_rate": 4.431821227552935e-05,
      "loss": 3.1534,
      "step": 61000
    },
    {
      "epoch": 0.67,
      "learning_rate": 4.426259715893862e-05,
      "loss": 3.1509,
      "step": 61500
    },
    {
      "epoch": 0.68,
      "learning_rate": 4.420675868846601e-05,
      "loss": 3.1336,
      "step": 62000
    },
    {
      "epoch": 0.68,
      "learning_rate": 4.415114357187528e-05,
      "loss": 3.1277,
      "step": 62500
    },
    {
      "epoch": 0.69,
      "learning_rate": 4.409530510140266e-05,
      "loss": 3.1124,
      "step": 63000
    },
    {
      "epoch": 0.69,
      "learning_rate": 4.403968998481194e-05,
      "loss": 3.1036,
      "step": 63500
    },
    {
      "epoch": 0.7,
      "learning_rate": 4.398385151433932e-05,
      "loss": 3.1021,
      "step": 64000
    },
    {
      "epoch": 0.7,
      "learning_rate": 4.3928236397748595e-05,
      "loss": 3.0982,
      "step": 64500
    },
    {
      "epoch": 0.71,
      "learning_rate": 4.387239792727598e-05,
      "loss": 3.0872,
      "step": 65000
    },
    {
      "epoch": 0.72,
      "learning_rate": 4.3816782810685253e-05,
      "loss": 3.0762,
      "step": 65500
    },
    {
      "epoch": 0.72,
      "learning_rate": 4.376094434021264e-05,
      "loss": 3.0731,
      "step": 66000
    },
    {
      "epoch": 0.73,
      "learning_rate": 4.370532922362191e-05,
      "loss": 3.0649,
      "step": 66500
    },
    {
      "epoch": 0.73,
      "learning_rate": 4.3649490753149295e-05,
      "loss": 3.0516,
      "step": 67000
    },
    {
      "epoch": 0.74,
      "learning_rate": 4.359387563655857e-05,
      "loss": 3.0461,
      "step": 67500
    },
    {
      "epoch": 0.74,
      "learning_rate": 4.3538037166085946e-05,
      "loss": 3.0406,
      "step": 68000
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.348242204949523e-05,
      "loss": 3.0346,
      "step": 68500
    },
    {
      "epoch": 0.75,
      "learning_rate": 4.3426583579022604e-05,
      "loss": 3.0305,
      "step": 69000
    },
    {
      "epoch": 0.76,
      "learning_rate": 4.337096846243188e-05,
      "loss": 3.0183,
      "step": 69500
    },
    {
      "epoch": 0.76,
      "learning_rate": 4.331512999195926e-05,
      "loss": 3.0115,
      "step": 70000
    },
    {
      "epoch": 0.76,
      "eval_accuracy": 0.4804065862365683,
      "eval_loss": 2.9140625,
      "eval_runtime": 1148.0843,
      "eval_samples_per_second": 268.583,
      "eval_steps_per_second": 4.197,
      "step": 70000
    },
    {
      "epoch": 0.77,
      "learning_rate": 4.325951487536854e-05,
      "loss": 3.0062,
      "step": 70500
    },
    {
      "epoch": 0.78,
      "learning_rate": 4.320367640489592e-05,
      "loss": 2.9965,
      "step": 71000
    },
    {
      "epoch": 0.78,
      "learning_rate": 4.3148061288305195e-05,
      "loss": 2.996,
      "step": 71500
    },
    {
      "epoch": 0.79,
      "learning_rate": 4.309222281783258e-05,
      "loss": 2.9863,
      "step": 72000
    },
    {
      "epoch": 0.79,
      "learning_rate": 4.3036607701241846e-05,
      "loss": 2.9806,
      "step": 72500
    },
    {
      "epoch": 0.8,
      "learning_rate": 4.2980769230769236e-05,
      "loss": 2.9721,
      "step": 73000
    },
    {
      "epoch": 0.8,
      "learning_rate": 4.2925154114178504e-05,
      "loss": 2.9673,
      "step": 73500
    },
    {
      "epoch": 0.81,
      "learning_rate": 4.286931564370589e-05,
      "loss": 2.9626,
      "step": 74000
    },
    {
      "epoch": 0.81,
      "learning_rate": 4.281370052711516e-05,
      "loss": 2.9566,
      "step": 74500
    },
    {
      "epoch": 0.82,
      "learning_rate": 4.2757862056642545e-05,
      "loss": 2.9434,
      "step": 75000
    },
    {
      "epoch": 0.82,
      "learning_rate": 4.270224694005182e-05,
      "loss": 2.9452,
      "step": 75500
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.2646408469579203e-05,
      "loss": 2.9393,
      "step": 76000
    },
    {
      "epoch": 0.84,
      "learning_rate": 4.259079335298848e-05,
      "loss": 2.9264,
      "step": 76500
    },
    {
      "epoch": 0.84,
      "learning_rate": 4.253495488251586e-05,
      "loss": 2.9295,
      "step": 77000
    },
    {
      "epoch": 0.85,
      "learning_rate": 4.2479339765925136e-05,
      "loss": 2.9244,
      "step": 77500
    },
    {
      "epoch": 0.85,
      "learning_rate": 4.242350129545252e-05,
      "loss": 2.9103,
      "step": 78000
    },
    {
      "epoch": 0.86,
      "learning_rate": 4.2367886178861794e-05,
      "loss": 2.909,
      "step": 78500
    },
    {
      "epoch": 0.86,
      "learning_rate": 4.231204770838918e-05,
      "loss": 2.9023,
      "step": 79000
    },
    {
      "epoch": 0.87,
      "learning_rate": 4.2256432591798446e-05,
      "loss": 2.897,
      "step": 79500
    },
    {
      "epoch": 0.87,
      "learning_rate": 4.220059412132583e-05,
      "loss": 2.8893,
      "step": 80000
    },
    {
      "epoch": 0.87,
      "eval_accuracy": 0.49558027459098025,
      "eval_loss": 2.79296875,
      "eval_runtime": 1161.4766,
      "eval_samples_per_second": 265.486,
      "eval_steps_per_second": 4.149,
      "step": 80000
    },
    {
      "epoch": 0.88,
      "learning_rate": 4.2144979004735104e-05,
      "loss": 2.8842,
      "step": 80500
    },
    {
      "epoch": 0.88,
      "learning_rate": 4.208914053426249e-05,
      "loss": 2.8806,
      "step": 81000
    },
    {
      "epoch": 0.89,
      "learning_rate": 4.203352541767176e-05,
      "loss": 2.878,
      "step": 81500
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.1977686947199145e-05,
      "loss": 2.866,
      "step": 82000
    },
    {
      "epoch": 0.9,
      "learning_rate": 4.192207183060842e-05,
      "loss": 2.868,
      "step": 82500
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.18662333601358e-05,
      "loss": 2.8598,
      "step": 83000
    },
    {
      "epoch": 0.91,
      "learning_rate": 4.181061824354508e-05,
      "loss": 2.8617,
      "step": 83500
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.1754779773072454e-05,
      "loss": 2.8433,
      "step": 84000
    },
    {
      "epoch": 0.92,
      "learning_rate": 4.169916465648173e-05,
      "loss": 2.8513,
      "step": 84500
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.164332618600911e-05,
      "loss": 2.8363,
      "step": 85000
    },
    {
      "epoch": 0.93,
      "learning_rate": 4.158771106941839e-05,
      "loss": 2.8388,
      "step": 85500
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.153187259894577e-05,
      "loss": 2.8346,
      "step": 86000
    },
    {
      "epoch": 0.94,
      "learning_rate": 4.1476257482355045e-05,
      "loss": 2.8296,
      "step": 86500
    },
    {
      "epoch": 0.95,
      "learning_rate": 4.142041901188243e-05,
      "loss": 2.8238,
      "step": 87000
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.13648038952917e-05,
      "loss": 2.8172,
      "step": 87500
    },
    {
      "epoch": 0.96,
      "learning_rate": 4.1308965424819086e-05,
      "loss": 2.8156,
      "step": 88000
    },
    {
      "epoch": 0.97,
      "learning_rate": 4.125335030822836e-05,
      "loss": 2.8063,
      "step": 88500
    },
    {
      "epoch": 0.97,
      "learning_rate": 4.1197511837755744e-05,
      "loss": 2.8078,
      "step": 89000
    },
    {
      "epoch": 0.98,
      "learning_rate": 4.114189672116502e-05,
      "loss": 2.7965,
      "step": 89500
    },
    {
      "epoch": 0.98,
      "learning_rate": 4.10860582506924e-05,
      "loss": 2.7983,
      "step": 90000
    },
    {
      "epoch": 0.98,
      "eval_accuracy": 0.5080531872849392,
      "eval_loss": 2.697265625,
      "eval_runtime": 1166.2832,
      "eval_samples_per_second": 264.392,
      "eval_steps_per_second": 4.132,
      "step": 90000
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.103044313410167e-05,
      "loss": 2.7939,
      "step": 90500
    },
    {
      "epoch": 0.99,
      "learning_rate": 4.097460466362906e-05,
      "loss": 2.785,
      "step": 91000
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.091898954703833e-05,
      "loss": 2.7767,
      "step": 91500
    },
    {
      "epoch": 1.0,
      "learning_rate": 4.086315107656571e-05,
      "loss": 2.7715,
      "step": 92000
    },
    {
      "epoch": 1.01,
      "learning_rate": 4.0807535959974986e-05,
      "loss": 2.7693,
      "step": 92500
    },
    {
      "epoch": 1.02,
      "learning_rate": 4.075169748950237e-05,
      "loss": 2.756,
      "step": 93000
    },
    {
      "epoch": 1.02,
      "learning_rate": 4.0696082372911645e-05,
      "loss": 2.7613,
      "step": 93500
    },
    {
      "epoch": 1.03,
      "learning_rate": 4.064024390243902e-05,
      "loss": 2.7611,
      "step": 94000
    },
    {
      "epoch": 1.03,
      "learning_rate": 4.05846287858483e-05,
      "loss": 2.7548,
      "step": 94500
    },
    {
      "epoch": 1.04,
      "learning_rate": 4.052879031537568e-05,
      "loss": 2.7462,
      "step": 95000
    },
    {
      "epoch": 1.04,
      "learning_rate": 4.047317519878496e-05,
      "loss": 2.7428,
      "step": 95500
    },
    {
      "epoch": 1.05,
      "learning_rate": 4.041733672831234e-05,
      "loss": 2.7396,
      "step": 96000
    },
    {
      "epoch": 1.05,
      "learning_rate": 4.036172161172161e-05,
      "loss": 2.7386,
      "step": 96500
    },
    {
      "epoch": 1.06,
      "learning_rate": 4.0305883141248995e-05,
      "loss": 2.7342,
      "step": 97000
    },
    {
      "epoch": 1.07,
      "learning_rate": 4.025026802465827e-05,
      "loss": 2.7322,
      "step": 97500
    },
    {
      "epoch": 1.07,
      "learning_rate": 4.019442955418565e-05,
      "loss": 2.7236,
      "step": 98000
    },
    {
      "epoch": 1.08,
      "learning_rate": 4.013881443759493e-05,
      "loss": 2.7217,
      "step": 98500
    },
    {
      "epoch": 1.08,
      "learning_rate": 4.008297596712231e-05,
      "loss": 2.709,
      "step": 99000
    },
    {
      "epoch": 1.09,
      "learning_rate": 4.0027360850531586e-05,
      "loss": 2.7157,
      "step": 99500
    },
    {
      "epoch": 1.09,
      "learning_rate": 3.997152238005897e-05,
      "loss": 2.7039,
      "step": 100000
    },
    {
      "epoch": 1.09,
      "eval_accuracy": 0.5214649672921134,
      "eval_loss": 2.6015625,
      "eval_runtime": 1419.9277,
      "eval_samples_per_second": 217.163,
      "eval_steps_per_second": 3.394,
      "step": 100000
    },
    {
      "epoch": 1.1,
      "learning_rate": 3.991590726346824e-05,
      "loss": 2.7006,
      "step": 100500
    },
    {
      "epoch": 1.1,
      "learning_rate": 3.986006879299563e-05,
      "loss": 2.6918,
      "step": 101000
    },
    {
      "epoch": 1.11,
      "learning_rate": 3.9804453676404895e-05,
      "loss": 2.6905,
      "step": 101500
    },
    {
      "epoch": 1.11,
      "learning_rate": 3.9748615205932285e-05,
      "loss": 2.6767,
      "step": 102000
    },
    {
      "epoch": 1.12,
      "learning_rate": 3.969300008934155e-05,
      "loss": 2.6684,
      "step": 102500
    },
    {
      "epoch": 1.13,
      "learning_rate": 3.963716161886894e-05,
      "loss": 2.6593,
      "step": 103000
    },
    {
      "epoch": 1.13,
      "learning_rate": 3.958154650227821e-05,
      "loss": 2.6453,
      "step": 103500
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.9525708031805594e-05,
      "loss": 2.6396,
      "step": 104000
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.947009291521487e-05,
      "loss": 2.6349,
      "step": 104500
    },
    {
      "epoch": 1.15,
      "learning_rate": 3.9414254444742246e-05,
      "loss": 2.6305,
      "step": 105000
    },
    {
      "epoch": 1.15,
      "learning_rate": 3.935863932815153e-05,
      "loss": 2.6234,
      "step": 105500
    },
    {
      "epoch": 1.16,
      "learning_rate": 3.9302800857678904e-05,
      "loss": 2.6105,
      "step": 106000
    },
    {
      "epoch": 1.16,
      "learning_rate": 3.9247185741088185e-05,
      "loss": 2.6056,
      "step": 106500
    },
    {
      "epoch": 1.17,
      "learning_rate": 3.919134727061556e-05,
      "loss": 2.6018,
      "step": 107000
    },
    {
      "epoch": 1.17,
      "learning_rate": 3.9135732154024843e-05,
      "loss": 2.5896,
      "step": 107500
    },
    {
      "epoch": 1.18,
      "learning_rate": 3.907989368355222e-05,
      "loss": 2.5921,
      "step": 108000
    },
    {
      "epoch": 1.19,
      "learning_rate": 3.9024278566961495e-05,
      "loss": 2.5853,
      "step": 108500
    },
    {
      "epoch": 1.19,
      "learning_rate": 3.896844009648888e-05,
      "loss": 2.5732,
      "step": 109000
    },
    {
      "epoch": 1.2,
      "learning_rate": 3.891282497989815e-05,
      "loss": 2.5678,
      "step": 109500
    },
    {
      "epoch": 1.2,
      "learning_rate": 3.8856986509425536e-05,
      "loss": 2.5658,
      "step": 110000
    },
    {
      "epoch": 1.2,
      "eval_accuracy": 0.5448239716997432,
      "eval_loss": 2.455078125,
      "eval_runtime": 1156.6399,
      "eval_samples_per_second": 266.596,
      "eval_steps_per_second": 4.166,
      "step": 110000
    },
    {
      "epoch": 1.21,
      "learning_rate": 3.8801371392834804e-05,
      "loss": 2.5615,
      "step": 110500
    },
    {
      "epoch": 1.21,
      "learning_rate": 3.8745532922362194e-05,
      "loss": 2.5516,
      "step": 111000
    },
    {
      "epoch": 1.22,
      "learning_rate": 3.868991780577146e-05,
      "loss": 2.5538,
      "step": 111500
    },
    {
      "epoch": 1.22,
      "learning_rate": 3.863407933529885e-05,
      "loss": 2.5485,
      "step": 112000
    },
    {
      "epoch": 1.23,
      "learning_rate": 3.857846421870812e-05,
      "loss": 2.5425,
      "step": 112500
    },
    {
      "epoch": 1.23,
      "learning_rate": 3.852262574823551e-05,
      "loss": 2.5413,
      "step": 113000
    },
    {
      "epoch": 1.24,
      "learning_rate": 3.846701063164478e-05,
      "loss": 2.533,
      "step": 113500
    },
    {
      "epoch": 1.25,
      "learning_rate": 3.841117216117217e-05,
      "loss": 2.5268,
      "step": 114000
    },
    {
      "epoch": 1.25,
      "learning_rate": 3.8355557044581436e-05,
      "loss": 2.5288,
      "step": 114500
    },
    {
      "epoch": 1.26,
      "learning_rate": 3.829971857410882e-05,
      "loss": 2.5193,
      "step": 115000
    },
    {
      "epoch": 1.26,
      "learning_rate": 3.8244103457518094e-05,
      "loss": 2.5148,
      "step": 115500
    },
    {
      "epoch": 1.27,
      "learning_rate": 3.818826498704548e-05,
      "loss": 2.5149,
      "step": 116000
    },
    {
      "epoch": 1.27,
      "learning_rate": 3.813264987045475e-05,
      "loss": 2.5089,
      "step": 116500
    },
    {
      "epoch": 1.28,
      "learning_rate": 3.807681139998213e-05,
      "loss": 2.5051,
      "step": 117000
    },
    {
      "epoch": 1.28,
      "learning_rate": 3.802119628339141e-05,
      "loss": 2.4994,
      "step": 117500
    },
    {
      "epoch": 1.29,
      "learning_rate": 3.7965357812918787e-05,
      "loss": 2.4975,
      "step": 118000
    },
    {
      "epoch": 1.29,
      "learning_rate": 3.790974269632807e-05,
      "loss": 2.4973,
      "step": 118500
    },
    {
      "epoch": 1.3,
      "learning_rate": 3.7853904225855445e-05,
      "loss": 2.49,
      "step": 119000
    },
    {
      "epoch": 1.31,
      "learning_rate": 3.7798289109264726e-05,
      "loss": 2.4893,
      "step": 119500
    },
    {
      "epoch": 1.31,
      "learning_rate": 3.77424506387921e-05,
      "loss": 2.4846,
      "step": 120000
    },
    {
      "epoch": 1.31,
      "eval_accuracy": 0.5575897132745044,
      "eval_loss": 2.373046875,
      "eval_runtime": 1158.688,
      "eval_samples_per_second": 266.125,
      "eval_steps_per_second": 4.159,
      "step": 120000
    },
    {
      "epoch": 1.32,
      "learning_rate": 3.768683552220138e-05,
      "loss": 2.4811,
      "step": 120500
    },
    {
      "epoch": 1.32,
      "learning_rate": 3.763099705172876e-05,
      "loss": 2.4738,
      "step": 121000
    },
    {
      "epoch": 1.33,
      "learning_rate": 3.757538193513803e-05,
      "loss": 2.4697,
      "step": 121500
    },
    {
      "epoch": 1.33,
      "learning_rate": 3.751954346466542e-05,
      "loss": 2.466,
      "step": 122000
    },
    {
      "epoch": 1.34,
      "learning_rate": 3.746392834807469e-05,
      "loss": 2.4581,
      "step": 122500
    },
    {
      "epoch": 1.34,
      "learning_rate": 3.740808987760208e-05,
      "loss": 2.4666,
      "step": 123000
    },
    {
      "epoch": 1.35,
      "learning_rate": 3.7352474761011345e-05,
      "loss": 2.4585,
      "step": 123500
    },
    {
      "epoch": 1.35,
      "learning_rate": 3.7296636290538735e-05,
      "loss": 2.4605,
      "step": 124000
    },
    {
      "epoch": 1.36,
      "learning_rate": 3.7241021173948e-05,
      "loss": 2.4512,
      "step": 124500
    },
    {
      "epoch": 1.37,
      "learning_rate": 3.718518270347539e-05,
      "loss": 2.4536,
      "step": 125000
    },
    {
      "epoch": 1.37,
      "learning_rate": 3.712956758688466e-05,
      "loss": 2.4465,
      "step": 125500
    },
    {
      "epoch": 1.38,
      "learning_rate": 3.7073729116412044e-05,
      "loss": 2.4468,
      "step": 126000
    },
    {
      "epoch": 1.38,
      "learning_rate": 3.701811399982132e-05,
      "loss": 2.4401,
      "step": 126500
    },
    {
      "epoch": 1.39,
      "learning_rate": 3.69622755293487e-05,
      "loss": 2.4435,
      "step": 127000
    },
    {
      "epoch": 1.39,
      "learning_rate": 3.690666041275798e-05,
      "loss": 2.4338,
      "step": 127500
    },
    {
      "epoch": 1.4,
      "learning_rate": 3.685082194228536e-05,
      "loss": 2.431,
      "step": 128000
    },
    {
      "epoch": 1.4,
      "learning_rate": 3.6795206825694635e-05,
      "loss": 2.4314,
      "step": 128500
    },
    {
      "epoch": 1.41,
      "learning_rate": 3.673936835522201e-05,
      "loss": 2.4304,
      "step": 129000
    },
    {
      "epoch": 1.41,
      "learning_rate": 3.668375323863129e-05,
      "loss": 2.4251,
      "step": 129500
    },
    {
      "epoch": 1.42,
      "learning_rate": 3.662791476815867e-05,
      "loss": 2.4284,
      "step": 130000
    },
    {
      "epoch": 1.42,
      "eval_accuracy": 0.5663108295591563,
      "eval_loss": 2.31640625,
      "eval_runtime": 1123.8513,
      "eval_samples_per_second": 274.374,
      "eval_steps_per_second": 4.288,
      "step": 130000
    },
    {
      "epoch": 1.43,
      "learning_rate": 3.657229965156795e-05,
      "loss": 2.4186,
      "step": 130500
    },
    {
      "epoch": 1.43,
      "learning_rate": 3.651646118109533e-05,
      "loss": 2.4175,
      "step": 131000
    },
    {
      "epoch": 1.44,
      "learning_rate": 3.64608460645046e-05,
      "loss": 2.419,
      "step": 131500
    },
    {
      "epoch": 1.44,
      "learning_rate": 3.6405007594031985e-05,
      "loss": 2.4102,
      "step": 132000
    },
    {
      "epoch": 1.45,
      "learning_rate": 3.634939247744126e-05,
      "loss": 2.4141,
      "step": 132500
    },
    {
      "epoch": 1.45,
      "learning_rate": 3.6293554006968643e-05,
      "loss": 2.4113,
      "step": 133000
    },
    {
      "epoch": 1.46,
      "learning_rate": 3.623793889037791e-05,
      "loss": 2.4101,
      "step": 133500
    },
    {
      "epoch": 1.46,
      "learning_rate": 3.61821004199053e-05,
      "loss": 2.4115,
      "step": 134000
    },
    {
      "epoch": 1.47,
      "learning_rate": 3.612648530331457e-05,
      "loss": 2.4031,
      "step": 134500
    },
    {
      "epoch": 1.47,
      "learning_rate": 3.607064683284196e-05,
      "loss": 2.4077,
      "step": 135000
    },
    {
      "epoch": 1.48,
      "learning_rate": 3.601503171625123e-05,
      "loss": 2.3965,
      "step": 135500
    },
    {
      "epoch": 1.49,
      "learning_rate": 3.595919324577861e-05,
      "loss": 2.3925,
      "step": 136000
    },
    {
      "epoch": 1.49,
      "learning_rate": 3.5903578129187886e-05,
      "loss": 2.3889,
      "step": 136500
    },
    {
      "epoch": 1.5,
      "learning_rate": 3.584773965871527e-05,
      "loss": 2.385,
      "step": 137000
    },
    {
      "epoch": 1.5,
      "learning_rate": 3.5792124542124544e-05,
      "loss": 2.3869,
      "step": 137500
    },
    {
      "epoch": 1.51,
      "learning_rate": 3.573628607165193e-05,
      "loss": 2.3868,
      "step": 138000
    },
    {
      "epoch": 1.51,
      "learning_rate": 3.56806709550612e-05,
      "loss": 2.3853,
      "step": 138500
    },
    {
      "epoch": 1.52,
      "learning_rate": 3.5624832484588585e-05,
      "loss": 2.3829,
      "step": 139000
    },
    {
      "epoch": 1.52,
      "learning_rate": 3.556921736799786e-05,
      "loss": 2.3826,
      "step": 139500
    },
    {
      "epoch": 1.53,
      "learning_rate": 3.551337889752524e-05,
      "loss": 2.3723,
      "step": 140000
    },
    {
      "epoch": 1.53,
      "eval_accuracy": 0.5726494894199156,
      "eval_loss": 2.2734375,
      "eval_runtime": 1199.6259,
      "eval_samples_per_second": 257.043,
      "eval_steps_per_second": 4.017,
      "step": 140000
    },
    {
      "epoch": 1.53,
      "learning_rate": 3.545776378093452e-05,
      "loss": 2.3741,
      "step": 140500
    },
    {
      "epoch": 1.54,
      "learning_rate": 3.5401925310461894e-05,
      "loss": 2.377,
      "step": 141000
    },
    {
      "epoch": 1.55,
      "learning_rate": 3.5346310193871176e-05,
      "loss": 2.3667,
      "step": 141500
    },
    {
      "epoch": 1.55,
      "learning_rate": 3.529047172339855e-05,
      "loss": 2.3684,
      "step": 142000
    },
    {
      "epoch": 1.56,
      "learning_rate": 3.523485660680783e-05,
      "loss": 2.3676,
      "step": 142500
    },
    {
      "epoch": 1.56,
      "learning_rate": 3.517901813633521e-05,
      "loss": 2.3665,
      "step": 143000
    },
    {
      "epoch": 1.57,
      "learning_rate": 3.5123403019744485e-05,
      "loss": 2.3671,
      "step": 143500
    },
    {
      "epoch": 1.57,
      "learning_rate": 3.506756454927187e-05,
      "loss": 2.3632,
      "step": 144000
    },
    {
      "epoch": 1.58,
      "learning_rate": 3.501194943268114e-05,
      "loss": 2.3629,
      "step": 144500
    },
    {
      "epoch": 1.58,
      "learning_rate": 3.4956110962208526e-05,
      "loss": 2.3608,
      "step": 145000
    },
    {
      "epoch": 1.59,
      "learning_rate": 3.4900495845617794e-05,
      "loss": 2.358,
      "step": 145500
    },
    {
      "epoch": 1.59,
      "learning_rate": 3.4844657375145184e-05,
      "loss": 2.3542,
      "step": 146000
    },
    {
      "epoch": 1.6,
      "learning_rate": 3.478904225855445e-05,
      "loss": 2.3491,
      "step": 146500
    },
    {
      "epoch": 1.61,
      "learning_rate": 3.4733203788081836e-05,
      "loss": 2.3547,
      "step": 147000
    },
    {
      "epoch": 1.61,
      "learning_rate": 3.467758867149111e-05,
      "loss": 2.342,
      "step": 147500
    },
    {
      "epoch": 1.62,
      "learning_rate": 3.4621750201018494e-05,
      "loss": 2.3466,
      "step": 148000
    },
    {
      "epoch": 1.62,
      "learning_rate": 3.456613508442777e-05,
      "loss": 2.3514,
      "step": 148500
    },
    {
      "epoch": 1.63,
      "learning_rate": 3.451029661395515e-05,
      "loss": 2.3442,
      "step": 149000
    },
    {
      "epoch": 1.63,
      "learning_rate": 3.4454681497364427e-05,
      "loss": 2.3381,
      "step": 149500
    },
    {
      "epoch": 1.64,
      "learning_rate": 3.439884302689181e-05,
      "loss": 2.3382,
      "step": 150000
    },
    {
      "epoch": 1.64,
      "eval_accuracy": 0.578659265524958,
      "eval_loss": 2.234375,
      "eval_runtime": 1153.4604,
      "eval_samples_per_second": 267.331,
      "eval_steps_per_second": 4.178,
      "step": 150000
    },
    {
      "epoch": 1.64,
      "learning_rate": 3.4343227910301085e-05,
      "loss": 2.3388,
      "step": 150500
    },
    {
      "epoch": 1.65,
      "learning_rate": 3.428738943982847e-05,
      "loss": 2.3338,
      "step": 151000
    },
    {
      "epoch": 1.65,
      "learning_rate": 3.423177432323774e-05,
      "loss": 2.3294,
      "step": 151500
    },
    {
      "epoch": 1.66,
      "learning_rate": 3.4175935852765126e-05,
      "loss": 2.3321,
      "step": 152000
    },
    {
      "epoch": 1.67,
      "learning_rate": 3.4120320736174394e-05,
      "loss": 2.3284,
      "step": 152500
    },
    {
      "epoch": 1.67,
      "learning_rate": 3.406448226570178e-05,
      "loss": 2.3289,
      "step": 153000
    },
    {
      "epoch": 1.68,
      "learning_rate": 3.400886714911105e-05,
      "loss": 2.3261,
      "step": 153500
    },
    {
      "epoch": 1.68,
      "learning_rate": 3.3953028678638435e-05,
      "loss": 2.3236,
      "step": 154000
    },
    {
      "epoch": 1.69,
      "learning_rate": 3.389741356204771e-05,
      "loss": 2.3273,
      "step": 154500
    },
    {
      "epoch": 1.69,
      "learning_rate": 3.384157509157509e-05,
      "loss": 2.3252,
      "step": 155000
    },
    {
      "epoch": 1.7,
      "learning_rate": 3.378595997498437e-05,
      "loss": 2.3183,
      "step": 155500
    },
    {
      "epoch": 1.7,
      "learning_rate": 3.373012150451175e-05,
      "loss": 2.3221,
      "step": 156000
    },
    {
      "epoch": 1.71,
      "learning_rate": 3.3674506387921026e-05,
      "loss": 2.3161,
      "step": 156500
    },
    {
      "epoch": 1.72,
      "learning_rate": 3.36186679174484e-05,
      "loss": 2.3116,
      "step": 157000
    },
    {
      "epoch": 1.72,
      "learning_rate": 3.356305280085768e-05,
      "loss": 2.3156,
      "step": 157500
    },
    {
      "epoch": 1.73,
      "learning_rate": 3.350721433038506e-05,
      "loss": 2.3084,
      "step": 158000
    },
    {
      "epoch": 1.73,
      "learning_rate": 3.3451599213794335e-05,
      "loss": 2.3133,
      "step": 158500
    },
    {
      "epoch": 1.74,
      "learning_rate": 3.339576074332172e-05,
      "loss": 2.3071,
      "step": 159000
    },
    {
      "epoch": 1.74,
      "learning_rate": 3.334014562673099e-05,
      "loss": 2.3062,
      "step": 159500
    },
    {
      "epoch": 1.75,
      "learning_rate": 3.3284307156258376e-05,
      "loss": 2.3084,
      "step": 160000
    },
    {
      "epoch": 1.75,
      "eval_accuracy": 0.5828903316963407,
      "eval_loss": 2.203125,
      "eval_runtime": 1108.8947,
      "eval_samples_per_second": 278.075,
      "eval_steps_per_second": 4.346,
      "step": 160000
    },
    {
      "epoch": 1.75,
      "learning_rate": 3.322869203966765e-05,
      "loss": 2.308,
      "step": 160500
    },
    {
      "epoch": 1.76,
      "learning_rate": 3.3172853569195035e-05,
      "loss": 2.3002,
      "step": 161000
    },
    {
      "epoch": 1.76,
      "learning_rate": 3.311723845260431e-05,
      "loss": 2.3019,
      "step": 161500
    },
    {
      "epoch": 1.77,
      "learning_rate": 3.306139998213169e-05,
      "loss": 2.3005,
      "step": 162000
    },
    {
      "epoch": 1.78,
      "learning_rate": 3.300578486554097e-05,
      "loss": 2.297,
      "step": 162500
    },
    {
      "epoch": 1.78,
      "learning_rate": 3.294994639506835e-05,
      "loss": 2.2935,
      "step": 163000
    },
    {
      "epoch": 1.79,
      "learning_rate": 3.289433127847762e-05,
      "loss": 2.2973,
      "step": 163500
    },
    {
      "epoch": 1.79,
      "learning_rate": 3.283849280800501e-05,
      "loss": 2.2976,
      "step": 164000
    },
    {
      "epoch": 1.8,
      "learning_rate": 3.278287769141428e-05,
      "loss": 2.2893,
      "step": 164500
    },
    {
      "epoch": 1.8,
      "learning_rate": 3.272703922094166e-05,
      "loss": 2.2884,
      "step": 165000
    },
    {
      "epoch": 1.81,
      "learning_rate": 3.2671424104350935e-05,
      "loss": 2.2901,
      "step": 165500
    },
    {
      "epoch": 1.81,
      "learning_rate": 3.261558563387832e-05,
      "loss": 2.2905,
      "step": 166000
    },
    {
      "epoch": 1.82,
      "learning_rate": 3.255997051728759e-05,
      "loss": 2.2869,
      "step": 166500
    },
    {
      "epoch": 1.82,
      "learning_rate": 3.2504132046814976e-05,
      "loss": 2.2831,
      "step": 167000
    },
    {
      "epoch": 1.83,
      "learning_rate": 3.244851693022425e-05,
      "loss": 2.2877,
      "step": 167500
    },
    {
      "epoch": 1.84,
      "learning_rate": 3.239267845975163e-05,
      "loss": 2.2818,
      "step": 168000
    },
    {
      "epoch": 1.84,
      "learning_rate": 3.233706334316091e-05,
      "loss": 2.2816,
      "step": 168500
    },
    {
      "epoch": 1.85,
      "learning_rate": 3.2281224872688285e-05,
      "loss": 2.2795,
      "step": 169000
    },
    {
      "epoch": 1.85,
      "learning_rate": 3.222560975609756e-05,
      "loss": 2.2815,
      "step": 169500
    },
    {
      "epoch": 1.86,
      "learning_rate": 3.216977128562494e-05,
      "loss": 2.2773,
      "step": 170000
    },
    {
      "epoch": 1.86,
      "eval_accuracy": 0.5871798298147288,
      "eval_loss": 2.17578125,
      "eval_runtime": 1172.3989,
      "eval_samples_per_second": 263.013,
      "eval_steps_per_second": 4.11,
      "step": 170000
    },
    {
      "epoch": 1.86,
      "learning_rate": 3.211415616903422e-05,
      "loss": 2.27,
      "step": 170500
    },
    {
      "epoch": 1.87,
      "learning_rate": 3.20583176985616e-05,
      "loss": 2.2723,
      "step": 171000
    },
    {
      "epoch": 1.87,
      "learning_rate": 3.2002702581970876e-05,
      "loss": 2.2735,
      "step": 171500
    },
    {
      "epoch": 1.88,
      "learning_rate": 3.194686411149826e-05,
      "loss": 2.2721,
      "step": 172000
    },
    {
      "epoch": 1.88,
      "learning_rate": 3.1891248994907534e-05,
      "loss": 2.2682,
      "step": 172500
    },
    {
      "epoch": 1.89,
      "learning_rate": 3.183541052443492e-05,
      "loss": 2.2682,
      "step": 173000
    },
    {
      "epoch": 1.9,
      "learning_rate": 3.1779795407844185e-05,
      "loss": 2.2682,
      "step": 173500
    },
    {
      "epoch": 1.9,
      "learning_rate": 3.1723956937371575e-05,
      "loss": 2.2679,
      "step": 174000
    },
    {
      "epoch": 1.91,
      "learning_rate": 3.1668341820780843e-05,
      "loss": 2.2609,
      "step": 174500
    },
    {
      "epoch": 1.91,
      "learning_rate": 3.1612503350308233e-05,
      "loss": 2.2611,
      "step": 175000
    },
    {
      "epoch": 1.92,
      "learning_rate": 3.15568882337175e-05,
      "loss": 2.2625,
      "step": 175500
    },
    {
      "epoch": 1.92,
      "learning_rate": 3.150104976324489e-05,
      "loss": 2.2589,
      "step": 176000
    },
    {
      "epoch": 1.93,
      "learning_rate": 3.144543464665416e-05,
      "loss": 2.2585,
      "step": 176500
    },
    {
      "epoch": 1.93,
      "learning_rate": 3.138959617618154e-05,
      "loss": 2.2527,
      "step": 177000
    },
    {
      "epoch": 1.94,
      "learning_rate": 3.133398105959082e-05,
      "loss": 2.2508,
      "step": 177500
    },
    {
      "epoch": 1.94,
      "learning_rate": 3.12781425891182e-05,
      "loss": 2.2526,
      "step": 178000
    },
    {
      "epoch": 1.95,
      "learning_rate": 3.1222527472527476e-05,
      "loss": 2.2454,
      "step": 178500
    },
    {
      "epoch": 1.96,
      "learning_rate": 3.116668900205485e-05,
      "loss": 2.2515,
      "step": 179000
    },
    {
      "epoch": 1.96,
      "learning_rate": 3.1111073885464134e-05,
      "loss": 2.2536,
      "step": 179500
    },
    {
      "epoch": 1.97,
      "learning_rate": 3.105523541499151e-05,
      "loss": 2.2492,
      "step": 180000
    },
    {
      "epoch": 1.97,
      "eval_accuracy": 0.5908707471657701,
      "eval_loss": 2.1484375,
      "eval_runtime": 1104.0527,
      "eval_samples_per_second": 279.295,
      "eval_steps_per_second": 4.365,
      "step": 180000
    },
    {
      "epoch": 1.97,
      "learning_rate": 3.099962029840079e-05,
      "loss": 2.2476,
      "step": 180500
    },
    {
      "epoch": 1.98,
      "learning_rate": 3.094378182792817e-05,
      "loss": 2.2394,
      "step": 181000
    },
    {
      "epoch": 1.98,
      "learning_rate": 3.088816671133744e-05,
      "loss": 2.2467,
      "step": 181500
    },
    {
      "epoch": 1.99,
      "learning_rate": 3.0832328240864826e-05,
      "loss": 2.2448,
      "step": 182000
    },
    {
      "epoch": 1.99,
      "learning_rate": 3.07767131242741e-05,
      "loss": 2.2393,
      "step": 182500
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.0720874653801484e-05,
      "loss": 2.2405,
      "step": 183000
    },
    {
      "epoch": 2.0,
      "learning_rate": 3.066525953721076e-05,
      "loss": 2.2345,
      "step": 183500
    },
    {
      "epoch": 2.01,
      "learning_rate": 3.060942106673814e-05,
      "loss": 2.2393,
      "step": 184000
    },
    {
      "epoch": 2.02,
      "learning_rate": 3.055380595014741e-05,
      "loss": 2.2374,
      "step": 184500
    },
    {
      "epoch": 2.02,
      "learning_rate": 3.04979674796748e-05,
      "loss": 2.2343,
      "step": 185000
    },
    {
      "epoch": 2.03,
      "learning_rate": 3.044235236308407e-05,
      "loss": 2.2304,
      "step": 185500
    },
    {
      "epoch": 2.03,
      "learning_rate": 3.0386513892611458e-05,
      "loss": 2.2328,
      "step": 186000
    },
    {
      "epoch": 2.04,
      "learning_rate": 3.0330898776020726e-05,
      "loss": 2.2348,
      "step": 186500
    },
    {
      "epoch": 2.04,
      "learning_rate": 3.0275060305548113e-05,
      "loss": 2.2267,
      "step": 187000
    },
    {
      "epoch": 2.05,
      "learning_rate": 3.0219445188957384e-05,
      "loss": 2.2261,
      "step": 187500
    },
    {
      "epoch": 2.05,
      "learning_rate": 3.0163606718484767e-05,
      "loss": 2.2244,
      "step": 188000
    },
    {
      "epoch": 2.06,
      "learning_rate": 3.0107991601894042e-05,
      "loss": 2.2262,
      "step": 188500
    },
    {
      "epoch": 2.06,
      "learning_rate": 3.0052153131421422e-05,
      "loss": 2.2225,
      "step": 189000
    },
    {
      "epoch": 2.07,
      "learning_rate": 2.99965380148307e-05,
      "loss": 2.2229,
      "step": 189500
    },
    {
      "epoch": 2.08,
      "learning_rate": 2.994069954435808e-05,
      "loss": 2.2261,
      "step": 190000
    },
    {
      "epoch": 2.08,
      "eval_accuracy": 0.5943238319324563,
      "eval_loss": 2.123046875,
      "eval_runtime": 1195.159,
      "eval_samples_per_second": 258.004,
      "eval_steps_per_second": 4.032,
      "step": 190000
    },
    {
      "epoch": 2.08,
      "learning_rate": 2.988508442776736e-05,
      "loss": 2.22,
      "step": 190500
    },
    {
      "epoch": 2.09,
      "learning_rate": 2.9829245957294738e-05,
      "loss": 2.2136,
      "step": 191000
    },
    {
      "epoch": 2.09,
      "learning_rate": 2.9773630840704013e-05,
      "loss": 2.2215,
      "step": 191500
    },
    {
      "epoch": 2.1,
      "learning_rate": 2.9717792370231396e-05,
      "loss": 2.211,
      "step": 192000
    },
    {
      "epoch": 2.1,
      "learning_rate": 2.966217725364067e-05,
      "loss": 2.2163,
      "step": 192500
    },
    {
      "epoch": 2.11,
      "learning_rate": 2.9606338783168054e-05,
      "loss": 2.2117,
      "step": 193000
    },
    {
      "epoch": 2.11,
      "learning_rate": 2.955072366657733e-05,
      "loss": 2.2179,
      "step": 193500
    },
    {
      "epoch": 2.12,
      "learning_rate": 2.949488519610471e-05,
      "loss": 2.2176,
      "step": 194000
    },
    {
      "epoch": 2.12,
      "learning_rate": 2.943927007951398e-05,
      "loss": 2.2149,
      "step": 194500
    },
    {
      "epoch": 2.13,
      "learning_rate": 2.9383431609041367e-05,
      "loss": 2.2128,
      "step": 195000
    },
    {
      "epoch": 2.14,
      "learning_rate": 2.932781649245064e-05,
      "loss": 2.2059,
      "step": 195500
    },
    {
      "epoch": 2.14,
      "learning_rate": 2.9271978021978025e-05,
      "loss": 2.2123,
      "step": 196000
    },
    {
      "epoch": 2.15,
      "learning_rate": 2.9216362905387296e-05,
      "loss": 2.2066,
      "step": 196500
    },
    {
      "epoch": 2.15,
      "learning_rate": 2.9160524434914683e-05,
      "loss": 2.2072,
      "step": 197000
    },
    {
      "epoch": 2.16,
      "learning_rate": 2.9104909318323954e-05,
      "loss": 2.2063,
      "step": 197500
    },
    {
      "epoch": 2.16,
      "learning_rate": 2.904907084785134e-05,
      "loss": 2.2054,
      "step": 198000
    },
    {
      "epoch": 2.17,
      "learning_rate": 2.899345573126061e-05,
      "loss": 2.2032,
      "step": 198500
    },
    {
      "epoch": 2.17,
      "learning_rate": 2.8937617260787992e-05,
      "loss": 2.1995,
      "step": 199000
    },
    {
      "epoch": 2.18,
      "learning_rate": 2.8882002144197267e-05,
      "loss": 2.2018,
      "step": 199500
    },
    {
      "epoch": 2.18,
      "learning_rate": 2.882616367372465e-05,
      "loss": 2.1961,
      "step": 200000
    },
    {
      "epoch": 2.18,
      "eval_accuracy": 0.5975533594798018,
      "eval_loss": 2.1015625,
      "eval_runtime": 1253.116,
      "eval_samples_per_second": 246.071,
      "eval_steps_per_second": 3.846,
      "step": 200000
    },
    {
      "epoch": 2.19,
      "learning_rate": 2.8770548557133925e-05,
      "loss": 2.2031,
      "step": 200500
    },
    {
      "epoch": 2.2,
      "learning_rate": 2.8714710086661305e-05,
      "loss": 2.2029,
      "step": 201000
    },
    {
      "epoch": 2.2,
      "learning_rate": 2.8659094970070583e-05,
      "loss": 2.1957,
      "step": 201500
    },
    {
      "epoch": 2.21,
      "learning_rate": 2.8603256499597963e-05,
      "loss": 2.2022,
      "step": 202000
    },
    {
      "epoch": 2.21,
      "learning_rate": 2.854764138300724e-05,
      "loss": 2.1946,
      "step": 202500
    },
    {
      "epoch": 2.22,
      "learning_rate": 2.849180291253462e-05,
      "loss": 2.1959,
      "step": 203000
    },
    {
      "epoch": 2.22,
      "learning_rate": 2.8436187795943896e-05,
      "loss": 2.1911,
      "step": 203500
    },
    {
      "epoch": 2.23,
      "learning_rate": 2.838034932547128e-05,
      "loss": 2.1922,
      "step": 204000
    },
    {
      "epoch": 2.23,
      "learning_rate": 2.832473420888055e-05,
      "loss": 2.1917,
      "step": 204500
    },
    {
      "epoch": 2.24,
      "learning_rate": 2.8268895738407937e-05,
      "loss": 2.1853,
      "step": 205000
    },
    {
      "epoch": 2.24,
      "learning_rate": 2.8213280621817205e-05,
      "loss": 2.189,
      "step": 205500
    },
    {
      "epoch": 2.25,
      "learning_rate": 2.8157442151344592e-05,
      "loss": 2.1869,
      "step": 206000
    },
    {
      "epoch": 2.26,
      "learning_rate": 2.8101827034753863e-05,
      "loss": 2.182,
      "step": 206500
    },
    {
      "epoch": 2.26,
      "learning_rate": 2.804598856428125e-05,
      "loss": 2.189,
      "step": 207000
    },
    {
      "epoch": 2.27,
      "learning_rate": 2.799037344769052e-05,
      "loss": 2.1849,
      "step": 207500
    },
    {
      "epoch": 2.27,
      "learning_rate": 2.7934534977217908e-05,
      "loss": 2.1859,
      "step": 208000
    },
    {
      "epoch": 2.28,
      "learning_rate": 2.787891986062718e-05,
      "loss": 2.1852,
      "step": 208500
    },
    {
      "epoch": 2.28,
      "learning_rate": 2.782308139015456e-05,
      "loss": 2.1819,
      "step": 209000
    },
    {
      "epoch": 2.29,
      "learning_rate": 2.7767466273563837e-05,
      "loss": 2.1806,
      "step": 209500
    },
    {
      "epoch": 2.29,
      "learning_rate": 2.7711627803091217e-05,
      "loss": 2.1838,
      "step": 210000
    },
    {
      "epoch": 2.29,
      "eval_accuracy": 0.6004074262116587,
      "eval_loss": 2.08203125,
      "eval_runtime": 1303.2811,
      "eval_samples_per_second": 236.6,
      "eval_steps_per_second": 3.698,
      "step": 210000
    },
    {
      "epoch": 2.3,
      "learning_rate": 2.7656012686500492e-05,
      "loss": 2.181,
      "step": 210500
    },
    {
      "epoch": 2.3,
      "learning_rate": 2.7600174216027875e-05,
      "loss": 2.1811,
      "step": 211000
    },
    {
      "epoch": 2.31,
      "learning_rate": 2.754455909943715e-05,
      "loss": 2.1749,
      "step": 211500
    },
    {
      "epoch": 2.32,
      "learning_rate": 2.7488720628964533e-05,
      "loss": 2.1756,
      "step": 212000
    },
    {
      "epoch": 2.32,
      "learning_rate": 2.7433105512373808e-05,
      "loss": 2.1799,
      "step": 212500
    },
    {
      "epoch": 2.33,
      "learning_rate": 2.7377267041901188e-05,
      "loss": 2.1769,
      "step": 213000
    },
    {
      "epoch": 2.33,
      "learning_rate": 2.7321651925310466e-05,
      "loss": 2.1757,
      "step": 213500
    },
    {
      "epoch": 2.34,
      "learning_rate": 2.7265813454837846e-05,
      "loss": 2.1682,
      "step": 214000
    },
    {
      "epoch": 2.34,
      "learning_rate": 2.7210198338247124e-05,
      "loss": 2.1728,
      "step": 214500
    },
    {
      "epoch": 2.35,
      "learning_rate": 2.7154359867774504e-05,
      "loss": 2.168,
      "step": 215000
    },
    {
      "epoch": 2.35,
      "learning_rate": 2.7098744751183775e-05,
      "loss": 2.1687,
      "step": 215500
    },
    {
      "epoch": 2.36,
      "learning_rate": 2.7042906280711162e-05,
      "loss": 2.1653,
      "step": 216000
    },
    {
      "epoch": 2.36,
      "learning_rate": 2.6987291164120433e-05,
      "loss": 2.166,
      "step": 216500
    },
    {
      "epoch": 2.37,
      "learning_rate": 2.693145269364782e-05,
      "loss": 2.1648,
      "step": 217000
    },
    {
      "epoch": 2.38,
      "learning_rate": 2.6875837577057088e-05,
      "loss": 2.1605,
      "step": 217500
    },
    {
      "epoch": 2.38,
      "learning_rate": 2.6819999106584475e-05,
      "loss": 2.1664,
      "step": 218000
    },
    {
      "epoch": 2.39,
      "learning_rate": 2.6764383989993746e-05,
      "loss": 2.1656,
      "step": 218500
    },
    {
      "epoch": 2.39,
      "learning_rate": 2.6708545519521133e-05,
      "loss": 2.1603,
      "step": 219000
    },
    {
      "epoch": 2.4,
      "learning_rate": 2.6652930402930404e-05,
      "loss": 2.1596,
      "step": 219500
    },
    {
      "epoch": 2.4,
      "learning_rate": 2.6597091932457784e-05,
      "loss": 2.164,
      "step": 220000
    },
    {
      "epoch": 2.4,
      "eval_accuracy": 0.6030687006414185,
      "eval_loss": 2.064453125,
      "eval_runtime": 1207.9573,
      "eval_samples_per_second": 255.271,
      "eval_steps_per_second": 3.989,
      "step": 220000
    },
    {
      "epoch": 2.41,
      "learning_rate": 2.6541476815867062e-05,
      "loss": 2.1569,
      "step": 220500
    },
    {
      "epoch": 2.41,
      "learning_rate": 2.6485638345394442e-05,
      "loss": 2.1631,
      "step": 221000
    },
    {
      "epoch": 2.42,
      "learning_rate": 2.643002322880372e-05,
      "loss": 2.1595,
      "step": 221500
    },
    {
      "epoch": 2.43,
      "learning_rate": 2.63741847583311e-05,
      "loss": 2.1522,
      "step": 222000
    },
    {
      "epoch": 2.43,
      "learning_rate": 2.6318569641740375e-05,
      "loss": 2.1537,
      "step": 222500
    },
    {
      "epoch": 2.44,
      "learning_rate": 2.6262731171267758e-05,
      "loss": 2.1528,
      "step": 223000
    },
    {
      "epoch": 2.44,
      "learning_rate": 2.6207116054677033e-05,
      "loss": 2.156,
      "step": 223500
    },
    {
      "epoch": 2.45,
      "learning_rate": 2.6151277584204416e-05,
      "loss": 2.1547,
      "step": 224000
    },
    {
      "epoch": 2.45,
      "learning_rate": 2.609566246761369e-05,
      "loss": 2.1529,
      "step": 224500
    },
    {
      "epoch": 2.46,
      "learning_rate": 2.603982399714107e-05,
      "loss": 2.1536,
      "step": 225000
    },
    {
      "epoch": 2.46,
      "learning_rate": 2.5984208880550342e-05,
      "loss": 2.146,
      "step": 225500
    },
    {
      "epoch": 2.47,
      "learning_rate": 2.592837041007773e-05,
      "loss": 2.15,
      "step": 226000
    },
    {
      "epoch": 2.47,
      "learning_rate": 2.5872755293487e-05,
      "loss": 2.1504,
      "step": 226500
    },
    {
      "epoch": 2.48,
      "learning_rate": 2.5816916823014387e-05,
      "loss": 2.1475,
      "step": 227000
    },
    {
      "epoch": 2.49,
      "learning_rate": 2.5761301706423658e-05,
      "loss": 2.1517,
      "step": 227500
    },
    {
      "epoch": 2.49,
      "learning_rate": 2.5705463235951045e-05,
      "loss": 2.1447,
      "step": 228000
    },
    {
      "epoch": 2.5,
      "learning_rate": 2.5649848119360316e-05,
      "loss": 2.1413,
      "step": 228500
    },
    {
      "epoch": 2.5,
      "learning_rate": 2.5594009648887703e-05,
      "loss": 2.1433,
      "step": 229000
    },
    {
      "epoch": 2.51,
      "learning_rate": 2.553839453229697e-05,
      "loss": 2.1476,
      "step": 229500
    },
    {
      "epoch": 2.51,
      "learning_rate": 2.5482556061824354e-05,
      "loss": 2.1456,
      "step": 230000
    },
    {
      "epoch": 2.51,
      "eval_accuracy": 0.6052384630896759,
      "eval_loss": 2.046875,
      "eval_runtime": 1225.601,
      "eval_samples_per_second": 251.596,
      "eval_steps_per_second": 3.932,
      "step": 230000
    },
    {
      "epoch": 2.52,
      "learning_rate": 2.542694094523363e-05,
      "loss": 2.144,
      "step": 230500
    },
    {
      "epoch": 2.52,
      "learning_rate": 2.5371102474761012e-05,
      "loss": 2.1397,
      "step": 231000
    },
    {
      "epoch": 2.53,
      "learning_rate": 2.5315487358170287e-05,
      "loss": 2.1469,
      "step": 231500
    },
    {
      "epoch": 2.53,
      "learning_rate": 2.5259648887697667e-05,
      "loss": 2.1434,
      "step": 232000
    },
    {
      "epoch": 2.54,
      "learning_rate": 2.5204033771106945e-05,
      "loss": 2.1388,
      "step": 232500
    },
    {
      "epoch": 2.55,
      "learning_rate": 2.5148195300634325e-05,
      "loss": 2.1403,
      "step": 233000
    },
    {
      "epoch": 2.55,
      "learning_rate": 2.5092580184043603e-05,
      "loss": 2.1349,
      "step": 233500
    },
    {
      "epoch": 2.56,
      "learning_rate": 2.5036741713570983e-05,
      "loss": 2.1363,
      "step": 234000
    },
    {
      "epoch": 2.56,
      "learning_rate": 2.4981126596980258e-05,
      "loss": 2.143,
      "step": 234500
    },
    {
      "epoch": 2.57,
      "learning_rate": 2.492528812650764e-05,
      "loss": 2.1419,
      "step": 235000
    },
    {
      "epoch": 2.57,
      "learning_rate": 2.4869673009916912e-05,
      "loss": 2.1412,
      "step": 235500
    },
    {
      "epoch": 2.58,
      "learning_rate": 2.48138345394443e-05,
      "loss": 2.1327,
      "step": 236000
    },
    {
      "epoch": 2.58,
      "learning_rate": 2.475821942285357e-05,
      "loss": 2.1315,
      "step": 236500
    },
    {
      "epoch": 2.59,
      "learning_rate": 2.4702380952380953e-05,
      "loss": 2.1338,
      "step": 237000
    },
    {
      "epoch": 2.59,
      "learning_rate": 2.4646765835790228e-05,
      "loss": 2.1345,
      "step": 237500
    },
    {
      "epoch": 2.6,
      "learning_rate": 2.4590927365317608e-05,
      "loss": 2.1344,
      "step": 238000
    },
    {
      "epoch": 2.61,
      "learning_rate": 2.4535312248726883e-05,
      "loss": 2.1287,
      "step": 238500
    },
    {
      "epoch": 2.61,
      "learning_rate": 2.4479473778254266e-05,
      "loss": 2.1346,
      "step": 239000
    },
    {
      "epoch": 2.62,
      "learning_rate": 2.442385866166354e-05,
      "loss": 2.1285,
      "step": 239500
    },
    {
      "epoch": 2.62,
      "learning_rate": 2.4368020191190924e-05,
      "loss": 2.1308,
      "step": 240000
    },
    {
      "epoch": 2.62,
      "eval_accuracy": 0.6080040884730269,
      "eval_loss": 2.029296875,
      "eval_runtime": 1742.4401,
      "eval_samples_per_second": 176.968,
      "eval_steps_per_second": 2.766,
      "step": 240000
    },
    {
      "epoch": 2.63,
      "learning_rate": 2.43124050746002e-05,
      "loss": 2.13,
      "step": 240500
    },
    {
      "epoch": 2.63,
      "learning_rate": 2.4256566604127582e-05,
      "loss": 2.1307,
      "step": 241000
    },
    {
      "epoch": 2.64,
      "learning_rate": 2.4200951487536854e-05,
      "loss": 2.1279,
      "step": 241500
    },
    {
      "epoch": 2.64,
      "learning_rate": 2.414511301706424e-05,
      "loss": 2.1236,
      "step": 242000
    },
    {
      "epoch": 2.65,
      "learning_rate": 2.408949790047351e-05,
      "loss": 2.12,
      "step": 242500
    },
    {
      "epoch": 2.65,
      "learning_rate": 2.4033659430000895e-05,
      "loss": 2.1255,
      "step": 243000
    },
    {
      "epoch": 2.66,
      "learning_rate": 2.397804431341017e-05,
      "loss": 2.1237,
      "step": 243500
    },
    {
      "epoch": 2.67,
      "learning_rate": 2.392220584293755e-05,
      "loss": 2.1213,
      "step": 244000
    },
    {
      "epoch": 2.67,
      "learning_rate": 2.3866590726346824e-05,
      "loss": 2.1227,
      "step": 244500
    },
    {
      "epoch": 2.68,
      "learning_rate": 2.3810752255874208e-05,
      "loss": 2.1262,
      "step": 245000
    },
    {
      "epoch": 2.68,
      "learning_rate": 2.3755137139283482e-05,
      "loss": 2.1176,
      "step": 245500
    },
    {
      "epoch": 2.69,
      "learning_rate": 2.3699298668810866e-05,
      "loss": 2.1184,
      "step": 246000
    },
    {
      "epoch": 2.69,
      "learning_rate": 2.364368355222014e-05,
      "loss": 2.1175,
      "step": 246500
    },
    {
      "epoch": 2.7,
      "learning_rate": 2.3587845081747524e-05,
      "loss": 2.122,
      "step": 247000
    },
    {
      "epoch": 2.7,
      "learning_rate": 2.3532229965156795e-05,
      "loss": 2.1119,
      "step": 247500
    },
    {
      "epoch": 2.71,
      "learning_rate": 2.347639149468418e-05,
      "loss": 2.1136,
      "step": 248000
    },
    {
      "epoch": 2.71,
      "learning_rate": 2.3420776378093453e-05,
      "loss": 2.1187,
      "step": 248500
    },
    {
      "epoch": 2.72,
      "learning_rate": 2.3364937907620836e-05,
      "loss": 2.1187,
      "step": 249000
    },
    {
      "epoch": 2.73,
      "learning_rate": 2.3309322791030108e-05,
      "loss": 2.1161,
      "step": 249500
    },
    {
      "epoch": 2.73,
      "learning_rate": 2.325348432055749e-05,
      "loss": 2.1161,
      "step": 250000
    },
    {
      "epoch": 2.73,
      "eval_accuracy": 0.6101299952350318,
      "eval_loss": 2.013671875,
      "eval_runtime": 1209.7923,
      "eval_samples_per_second": 254.883,
      "eval_steps_per_second": 3.983,
      "step": 250000
    },
    {
      "epoch": 2.74,
      "learning_rate": 2.3197645850084877e-05,
      "loss": 2.1108,
      "step": 250500
    },
    {
      "epoch": 2.74,
      "learning_rate": 2.314203073349415e-05,
      "loss": 2.1159,
      "step": 251000
    },
    {
      "epoch": 2.75,
      "learning_rate": 2.3086192263021532e-05,
      "loss": 2.1105,
      "step": 251500
    },
    {
      "epoch": 2.75,
      "learning_rate": 2.3030577146430807e-05,
      "loss": 2.1081,
      "step": 252000
    },
    {
      "epoch": 2.76,
      "learning_rate": 2.297473867595819e-05,
      "loss": 2.1085,
      "step": 252500
    },
    {
      "epoch": 2.76,
      "learning_rate": 2.2919123559367465e-05,
      "loss": 2.1148,
      "step": 253000
    },
    {
      "epoch": 2.77,
      "learning_rate": 2.2863285088894845e-05,
      "loss": 2.1079,
      "step": 253500
    },
    {
      "epoch": 2.77,
      "learning_rate": 2.280766997230412e-05,
      "loss": 2.113,
      "step": 254000
    },
    {
      "epoch": 2.78,
      "learning_rate": 2.2751831501831503e-05,
      "loss": 2.1102,
      "step": 254500
    },
    {
      "epoch": 2.79,
      "learning_rate": 2.2696216385240778e-05,
      "loss": 2.1081,
      "step": 255000
    },
    {
      "epoch": 2.79,
      "learning_rate": 2.264037791476816e-05,
      "loss": 2.1121,
      "step": 255500
    },
    {
      "epoch": 2.8,
      "learning_rate": 2.2584762798177432e-05,
      "loss": 2.1069,
      "step": 256000
    },
    {
      "epoch": 2.8,
      "learning_rate": 2.252892432770482e-05,
      "loss": 2.105,
      "step": 256500
    },
    {
      "epoch": 2.81,
      "learning_rate": 2.247330921111409e-05,
      "loss": 2.1047,
      "step": 257000
    },
    {
      "epoch": 2.81,
      "learning_rate": 2.2417470740641474e-05,
      "loss": 2.1006,
      "step": 257500
    },
    {
      "epoch": 2.82,
      "learning_rate": 2.236185562405075e-05,
      "loss": 2.1086,
      "step": 258000
    },
    {
      "epoch": 2.82,
      "learning_rate": 2.2306017153578128e-05,
      "loss": 2.0998,
      "step": 258500
    },
    {
      "epoch": 2.83,
      "learning_rate": 2.2250402036987403e-05,
      "loss": 2.1076,
      "step": 259000
    },
    {
      "epoch": 2.83,
      "learning_rate": 2.2194563566514786e-05,
      "loss": 2.1039,
      "step": 259500
    },
    {
      "epoch": 2.84,
      "learning_rate": 2.213894844992406e-05,
      "loss": 2.1052,
      "step": 260000
    },
    {
      "epoch": 2.84,
      "eval_accuracy": 0.6120208554253012,
      "eval_loss": 2.001953125,
      "eval_runtime": 1054.4543,
      "eval_samples_per_second": 292.432,
      "eval_steps_per_second": 4.57,
      "step": 260000
    },
    {
      "epoch": 2.85,
      "learning_rate": 2.2083109979451444e-05,
      "loss": 2.098,
      "step": 260500
    },
    {
      "epoch": 2.85,
      "learning_rate": 2.202749486286072e-05,
      "loss": 2.0984,
      "step": 261000
    },
    {
      "epoch": 2.86,
      "learning_rate": 2.1971656392388102e-05,
      "loss": 2.1033,
      "step": 261500
    },
    {
      "epoch": 2.86,
      "learning_rate": 2.1916041275797374e-05,
      "loss": 2.0997,
      "step": 262000
    },
    {
      "epoch": 2.87,
      "learning_rate": 2.186020280532476e-05,
      "loss": 2.0943,
      "step": 262500
    },
    {
      "epoch": 2.87,
      "learning_rate": 2.1804587688734032e-05,
      "loss": 2.0861,
      "step": 263000
    },
    {
      "epoch": 2.88,
      "learning_rate": 2.1748749218261415e-05,
      "loss": 2.0976,
      "step": 263500
    },
    {
      "epoch": 2.88,
      "learning_rate": 2.1693134101670686e-05,
      "loss": 2.0934,
      "step": 264000
    },
    {
      "epoch": 2.89,
      "learning_rate": 2.163729563119807e-05,
      "loss": 2.0899,
      "step": 264500
    },
    {
      "epoch": 2.89,
      "learning_rate": 2.1581680514607344e-05,
      "loss": 2.0975,
      "step": 265000
    },
    {
      "epoch": 2.9,
      "learning_rate": 2.1525842044134728e-05,
      "loss": 2.0973,
      "step": 265500
    },
    {
      "epoch": 2.91,
      "learning_rate": 2.1470226927544002e-05,
      "loss": 2.0963,
      "step": 266000
    },
    {
      "epoch": 2.91,
      "learning_rate": 2.1414388457071386e-05,
      "loss": 2.0907,
      "step": 266500
    },
    {
      "epoch": 2.92,
      "learning_rate": 2.135877334048066e-05,
      "loss": 2.0895,
      "step": 267000
    },
    {
      "epoch": 2.92,
      "learning_rate": 2.1302934870008044e-05,
      "loss": 2.088,
      "step": 267500
    },
    {
      "epoch": 2.93,
      "learning_rate": 2.1247319753417315e-05,
      "loss": 2.087,
      "step": 268000
    },
    {
      "epoch": 2.93,
      "learning_rate": 2.11914812829447e-05,
      "loss": 2.0894,
      "step": 268500
    },
    {
      "epoch": 2.94,
      "learning_rate": 2.113586616635397e-05,
      "loss": 2.0888,
      "step": 269000
    },
    {
      "epoch": 2.94,
      "learning_rate": 2.1080027695881356e-05,
      "loss": 2.0838,
      "step": 269500
    },
    {
      "epoch": 2.95,
      "learning_rate": 2.1024412579290628e-05,
      "loss": 2.0856,
      "step": 270000
    },
    {
      "epoch": 2.95,
      "eval_accuracy": 0.6142498653012093,
      "eval_loss": 1.990234375,
      "eval_runtime": 1044.3766,
      "eval_samples_per_second": 295.254,
      "eval_steps_per_second": 4.614,
      "step": 270000
    },
    {
      "epoch": 2.95,
      "learning_rate": 2.096857410881801e-05,
      "loss": 2.0933,
      "step": 270500
    },
    {
      "epoch": 2.96,
      "learning_rate": 2.0912958992227286e-05,
      "loss": 2.0858,
      "step": 271000
    },
    {
      "epoch": 2.97,
      "learning_rate": 2.085712052175467e-05,
      "loss": 2.0825,
      "step": 271500
    },
    {
      "epoch": 2.97,
      "learning_rate": 2.0801505405163944e-05,
      "loss": 2.0779,
      "step": 272000
    },
    {
      "epoch": 2.98,
      "learning_rate": 2.0745666934691327e-05,
      "loss": 2.078,
      "step": 272500
    },
    {
      "epoch": 2.98,
      "learning_rate": 2.0690051818100602e-05,
      "loss": 2.0848,
      "step": 273000
    },
    {
      "epoch": 2.99,
      "learning_rate": 2.0634213347627982e-05,
      "loss": 2.0816,
      "step": 273500
    },
    {
      "epoch": 2.99,
      "learning_rate": 2.0578598231037257e-05,
      "loss": 2.0816,
      "step": 274000
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.052275976056464e-05,
      "loss": 2.0815,
      "step": 274500
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.046714464397391e-05,
      "loss": 2.0842,
      "step": 275000
    },
    {
      "epoch": 3.01,
      "learning_rate": 2.0411306173501298e-05,
      "loss": 2.0743,
      "step": 275500
    },
    {
      "epoch": 3.01,
      "learning_rate": 2.035569105691057e-05,
      "loss": 2.0809,
      "step": 276000
    },
    {
      "epoch": 3.02,
      "learning_rate": 2.0299852586437952e-05,
      "loss": 2.078,
      "step": 276500
    },
    {
      "epoch": 3.03,
      "learning_rate": 2.0244237469847227e-05,
      "loss": 2.0771,
      "step": 277000
    },
    {
      "epoch": 3.03,
      "learning_rate": 2.018839899937461e-05,
      "loss": 2.0744,
      "step": 277500
    },
    {
      "epoch": 3.04,
      "learning_rate": 2.0132783882783885e-05,
      "loss": 2.0766,
      "step": 278000
    },
    {
      "epoch": 3.04,
      "learning_rate": 2.007694541231127e-05,
      "loss": 2.074,
      "step": 278500
    },
    {
      "epoch": 3.05,
      "learning_rate": 2.0021330295720543e-05,
      "loss": 2.0783,
      "step": 279000
    },
    {
      "epoch": 3.05,
      "learning_rate": 1.9965491825247923e-05,
      "loss": 2.0759,
      "step": 279500
    },
    {
      "epoch": 3.06,
      "learning_rate": 1.9909876708657198e-05,
      "loss": 2.0743,
      "step": 280000
    },
    {
      "epoch": 3.06,
      "eval_accuracy": 0.6159319542506481,
      "eval_loss": 1.9775390625,
      "eval_runtime": 1055.3371,
      "eval_samples_per_second": 292.187,
      "eval_steps_per_second": 4.566,
      "step": 280000
    },
    {
      "epoch": 3.06,
      "learning_rate": 1.985403823818458e-05,
      "loss": 2.0781,
      "step": 280500
    },
    {
      "epoch": 3.07,
      "learning_rate": 1.9798423121593853e-05,
      "loss": 2.0757,
      "step": 281000
    },
    {
      "epoch": 3.08,
      "learning_rate": 1.974258465112124e-05,
      "loss": 2.0722,
      "step": 281500
    },
    {
      "epoch": 3.08,
      "learning_rate": 1.968696953453051e-05,
      "loss": 2.0736,
      "step": 282000
    },
    {
      "epoch": 3.09,
      "learning_rate": 1.9631131064057894e-05,
      "loss": 2.064,
      "step": 282500
    },
    {
      "epoch": 3.09,
      "learning_rate": 1.957551594746717e-05,
      "loss": 2.0652,
      "step": 283000
    },
    {
      "epoch": 3.1,
      "learning_rate": 1.9519677476994552e-05,
      "loss": 2.0628,
      "step": 283500
    },
    {
      "epoch": 3.1,
      "learning_rate": 1.9464062360403827e-05,
      "loss": 2.069,
      "step": 284000
    },
    {
      "epoch": 3.11,
      "learning_rate": 1.9408223889931207e-05,
      "loss": 2.0627,
      "step": 284500
    },
    {
      "epoch": 3.11,
      "learning_rate": 1.935260877334048e-05,
      "loss": 2.0675,
      "step": 285000
    },
    {
      "epoch": 3.12,
      "learning_rate": 1.9296770302867865e-05,
      "loss": 2.0697,
      "step": 285500
    },
    {
      "epoch": 3.12,
      "learning_rate": 1.924115518627714e-05,
      "loss": 2.0659,
      "step": 286000
    },
    {
      "epoch": 3.13,
      "learning_rate": 1.9185316715804523e-05,
      "loss": 2.0704,
      "step": 286500
    },
    {
      "epoch": 3.14,
      "learning_rate": 1.9129701599213794e-05,
      "loss": 2.0663,
      "step": 287000
    },
    {
      "epoch": 3.14,
      "learning_rate": 1.907386312874118e-05,
      "loss": 2.0673,
      "step": 287500
    },
    {
      "epoch": 3.15,
      "learning_rate": 1.9018248012150452e-05,
      "loss": 2.0637,
      "step": 288000
    },
    {
      "epoch": 3.15,
      "learning_rate": 1.8962409541677835e-05,
      "loss": 2.0562,
      "step": 288500
    },
    {
      "epoch": 3.16,
      "learning_rate": 1.890679442508711e-05,
      "loss": 2.0616,
      "step": 289000
    },
    {
      "epoch": 3.16,
      "learning_rate": 1.885095595461449e-05,
      "loss": 2.0647,
      "step": 289500
    },
    {
      "epoch": 3.17,
      "learning_rate": 1.8795340838023765e-05,
      "loss": 2.0598,
      "step": 290000
    },
    {
      "epoch": 3.17,
      "eval_accuracy": 0.6171213473795026,
      "eval_loss": 1.9677734375,
      "eval_runtime": 1057.101,
      "eval_samples_per_second": 291.7,
      "eval_steps_per_second": 4.559,
      "step": 290000
    },
    {
      "epoch": 3.17,
      "learning_rate": 1.8739502367551148e-05,
      "loss": 2.0678,
      "step": 290500
    },
    {
      "epoch": 3.18,
      "learning_rate": 1.8683887250960423e-05,
      "loss": 2.0628,
      "step": 291000
    },
    {
      "epoch": 3.18,
      "learning_rate": 1.8628048780487806e-05,
      "loss": 2.0625,
      "step": 291500
    },
    {
      "epoch": 3.19,
      "learning_rate": 1.857243366389708e-05,
      "loss": 2.0597,
      "step": 292000
    },
    {
      "epoch": 3.2,
      "learning_rate": 1.8516595193424464e-05,
      "loss": 2.054,
      "step": 292500
    },
    {
      "epoch": 3.2,
      "learning_rate": 1.8460980076833735e-05,
      "loss": 2.0521,
      "step": 293000
    },
    {
      "epoch": 3.21,
      "learning_rate": 1.8405141606361122e-05,
      "loss": 2.0619,
      "step": 293500
    },
    {
      "epoch": 3.21,
      "learning_rate": 1.8349526489770393e-05,
      "loss": 2.0588,
      "step": 294000
    },
    {
      "epoch": 3.22,
      "learning_rate": 1.8293688019297773e-05,
      "loss": 2.0538,
      "step": 294500
    },
    {
      "epoch": 3.22,
      "learning_rate": 1.823807290270705e-05,
      "loss": 2.0571,
      "step": 295000
    },
    {
      "epoch": 3.23,
      "learning_rate": 1.818223443223443e-05,
      "loss": 2.0547,
      "step": 295500
    },
    {
      "epoch": 3.23,
      "learning_rate": 1.8126619315643706e-05,
      "loss": 2.0563,
      "step": 296000
    },
    {
      "epoch": 3.24,
      "learning_rate": 1.807078084517109e-05,
      "loss": 2.0484,
      "step": 296500
    },
    {
      "epoch": 3.24,
      "learning_rate": 1.8015165728580364e-05,
      "loss": 2.0563,
      "step": 297000
    },
    {
      "epoch": 3.25,
      "learning_rate": 1.7959327258107747e-05,
      "loss": 2.0537,
      "step": 297500
    },
    {
      "epoch": 3.26,
      "learning_rate": 1.7903712141517022e-05,
      "loss": 2.0554,
      "step": 298000
    },
    {
      "epoch": 3.26,
      "learning_rate": 1.7847873671044405e-05,
      "loss": 2.0508,
      "step": 298500
    },
    {
      "epoch": 3.27,
      "learning_rate": 1.7792258554453677e-05,
      "loss": 2.0515,
      "step": 299000
    },
    {
      "epoch": 3.27,
      "learning_rate": 1.7736420083981063e-05,
      "loss": 2.0529,
      "step": 299500
    },
    {
      "epoch": 3.28,
      "learning_rate": 1.7680804967390335e-05,
      "loss": 2.0492,
      "step": 300000
    },
    {
      "epoch": 3.28,
      "eval_accuracy": 0.6189548216499007,
      "eval_loss": 1.9560546875,
      "eval_runtime": 1033.7042,
      "eval_samples_per_second": 298.302,
      "eval_steps_per_second": 4.662,
      "step": 300000
    },
    {
      "epoch": 3.28,
      "learning_rate": 1.7624966496917715e-05,
      "loss": 2.0521,
      "step": 300500
    },
    {
      "epoch": 3.29,
      "learning_rate": 1.756935138032699e-05,
      "loss": 2.049,
      "step": 301000
    },
    {
      "epoch": 3.29,
      "learning_rate": 1.7513512909854373e-05,
      "loss": 2.0561,
      "step": 301500
    },
    {
      "epoch": 3.3,
      "learning_rate": 1.7457897793263648e-05,
      "loss": 2.0503,
      "step": 302000
    },
    {
      "epoch": 3.3,
      "learning_rate": 1.740205932279103e-05,
      "loss": 2.0487,
      "step": 302500
    },
    {
      "epoch": 3.31,
      "learning_rate": 1.7346444206200306e-05,
      "loss": 2.0489,
      "step": 303000
    },
    {
      "epoch": 3.32,
      "learning_rate": 1.729060573572769e-05,
      "loss": 2.0521,
      "step": 303500
    },
    {
      "epoch": 3.32,
      "learning_rate": 1.7234990619136964e-05,
      "loss": 2.0519,
      "step": 304000
    },
    {
      "epoch": 3.33,
      "learning_rate": 1.7179152148664347e-05,
      "loss": 2.0493,
      "step": 304500
    },
    {
      "epoch": 3.33,
      "learning_rate": 1.7123537032073618e-05,
      "loss": 2.0487,
      "step": 305000
    },
    {
      "epoch": 3.34,
      "learning_rate": 1.7067698561601e-05,
      "loss": 2.0428,
      "step": 305500
    },
    {
      "epoch": 3.34,
      "learning_rate": 1.7012083445010273e-05,
      "loss": 2.0482,
      "step": 306000
    },
    {
      "epoch": 3.35,
      "learning_rate": 1.6956244974537656e-05,
      "loss": 2.0427,
      "step": 306500
    },
    {
      "epoch": 3.35,
      "learning_rate": 1.690062985794693e-05,
      "loss": 2.0488,
      "step": 307000
    },
    {
      "epoch": 3.36,
      "learning_rate": 1.6844791387474314e-05,
      "loss": 2.0423,
      "step": 307500
    },
    {
      "epoch": 3.36,
      "learning_rate": 1.678917627088359e-05,
      "loss": 2.0426,
      "step": 308000
    },
    {
      "epoch": 3.37,
      "learning_rate": 1.6733337800410972e-05,
      "loss": 2.0402,
      "step": 308500
    },
    {
      "epoch": 3.38,
      "learning_rate": 1.6677722683820247e-05,
      "loss": 2.0465,
      "step": 309000
    },
    {
      "epoch": 3.38,
      "learning_rate": 1.662188421334763e-05,
      "loss": 2.0397,
      "step": 309500
    },
    {
      "epoch": 3.39,
      "learning_rate": 1.6566269096756905e-05,
      "loss": 2.0395,
      "step": 310000
    },
    {
      "epoch": 3.39,
      "eval_accuracy": 0.6202979048164721,
      "eval_loss": 1.9453125,
      "eval_runtime": 1032.8224,
      "eval_samples_per_second": 298.557,
      "eval_steps_per_second": 4.666,
      "step": 310000
    },
    {
      "epoch": 3.39,
      "learning_rate": 1.6510430626284285e-05,
      "loss": 2.0398,
      "step": 310500
    },
    {
      "epoch": 3.4,
      "learning_rate": 1.6454815509693556e-05,
      "loss": 2.0433,
      "step": 311000
    },
    {
      "epoch": 3.4,
      "learning_rate": 1.6398977039220943e-05,
      "loss": 2.0405,
      "step": 311500
    },
    {
      "epoch": 3.41,
      "learning_rate": 1.6343361922630214e-05,
      "loss": 2.0379,
      "step": 312000
    },
    {
      "epoch": 3.41,
      "learning_rate": 1.6287523452157598e-05,
      "loss": 2.0427,
      "step": 312500
    },
    {
      "epoch": 3.42,
      "learning_rate": 1.6231908335566872e-05,
      "loss": 2.0372,
      "step": 313000
    },
    {
      "epoch": 3.42,
      "learning_rate": 1.6176069865094256e-05,
      "loss": 2.0357,
      "step": 313500
    },
    {
      "epoch": 3.43,
      "learning_rate": 1.612045474850353e-05,
      "loss": 2.0375,
      "step": 314000
    },
    {
      "epoch": 3.44,
      "learning_rate": 1.6064616278030914e-05,
      "loss": 2.0395,
      "step": 314500
    },
    {
      "epoch": 3.44,
      "learning_rate": 1.600900116144019e-05,
      "loss": 2.0443,
      "step": 315000
    },
    {
      "epoch": 3.45,
      "learning_rate": 1.5953162690967568e-05,
      "loss": 2.0344,
      "step": 315500
    },
    {
      "epoch": 3.45,
      "learning_rate": 1.5897547574376843e-05,
      "loss": 2.0369,
      "step": 316000
    },
    {
      "epoch": 3.46,
      "learning_rate": 1.5841709103904226e-05,
      "loss": 2.0364,
      "step": 316500
    },
    {
      "epoch": 3.46,
      "learning_rate": 1.5786093987313498e-05,
      "loss": 2.0337,
      "step": 317000
    },
    {
      "epoch": 3.47,
      "learning_rate": 1.5730255516840884e-05,
      "loss": 2.0381,
      "step": 317500
    },
    {
      "epoch": 3.47,
      "learning_rate": 1.5674640400250156e-05,
      "loss": 2.0333,
      "step": 318000
    },
    {
      "epoch": 3.48,
      "learning_rate": 1.561880192977754e-05,
      "loss": 2.0311,
      "step": 318500
    },
    {
      "epoch": 3.48,
      "learning_rate": 1.5563186813186814e-05,
      "loss": 2.0311,
      "step": 319000
    },
    {
      "epoch": 3.49,
      "learning_rate": 1.5507348342714197e-05,
      "loss": 2.0287,
      "step": 319500
    },
    {
      "epoch": 3.5,
      "learning_rate": 1.5451733226123472e-05,
      "loss": 2.0328,
      "step": 320000
    },
    {
      "epoch": 3.5,
      "eval_accuracy": 0.6217200609754877,
      "eval_loss": 1.9365234375,
      "eval_runtime": 1035.5409,
      "eval_samples_per_second": 297.773,
      "eval_steps_per_second": 4.654,
      "step": 320000
    },
    {
      "epoch": 3.5,
      "learning_rate": 1.539589475565085e-05,
      "loss": 2.0322,
      "step": 320500
    },
    {
      "epoch": 3.51,
      "learning_rate": 1.534027963906013e-05,
      "loss": 2.0326,
      "step": 321000
    },
    {
      "epoch": 3.51,
      "learning_rate": 1.528444116858751e-05,
      "loss": 2.033,
      "step": 321500
    },
    {
      "epoch": 3.52,
      "learning_rate": 1.5228826051996783e-05,
      "loss": 2.0315,
      "step": 322000
    },
    {
      "epoch": 3.52,
      "learning_rate": 1.5172987581524168e-05,
      "loss": 2.0346,
      "step": 322500
    },
    {
      "epoch": 3.53,
      "learning_rate": 1.511737246493344e-05,
      "loss": 2.0256,
      "step": 323000
    },
    {
      "epoch": 3.53,
      "learning_rate": 1.5061533994460824e-05,
      "loss": 2.0242,
      "step": 323500
    },
    {
      "epoch": 3.54,
      "learning_rate": 1.5005918877870099e-05,
      "loss": 2.028,
      "step": 324000
    },
    {
      "epoch": 3.54,
      "learning_rate": 1.4950080407397482e-05,
      "loss": 2.0291,
      "step": 324500
    },
    {
      "epoch": 3.55,
      "learning_rate": 1.4894465290806755e-05,
      "loss": 2.0266,
      "step": 325000
    },
    {
      "epoch": 3.56,
      "learning_rate": 1.483862682033414e-05,
      "loss": 2.027,
      "step": 325500
    },
    {
      "epoch": 3.56,
      "learning_rate": 1.4783011703743413e-05,
      "loss": 2.0302,
      "step": 326000
    },
    {
      "epoch": 3.57,
      "learning_rate": 1.4727173233270795e-05,
      "loss": 2.0297,
      "step": 326500
    },
    {
      "epoch": 3.57,
      "learning_rate": 1.4671558116680068e-05,
      "loss": 2.0317,
      "step": 327000
    },
    {
      "epoch": 3.58,
      "learning_rate": 1.4615719646207451e-05,
      "loss": 2.0252,
      "step": 327500
    },
    {
      "epoch": 3.58,
      "learning_rate": 1.4560104529616724e-05,
      "loss": 2.0277,
      "step": 328000
    },
    {
      "epoch": 3.59,
      "learning_rate": 1.4504266059144109e-05,
      "loss": 2.0303,
      "step": 328500
    },
    {
      "epoch": 3.59,
      "learning_rate": 1.4448650942553382e-05,
      "loss": 2.0247,
      "step": 329000
    },
    {
      "epoch": 3.6,
      "learning_rate": 1.4392812472080765e-05,
      "loss": 2.0228,
      "step": 329500
    },
    {
      "epoch": 3.6,
      "learning_rate": 1.433719735549004e-05,
      "loss": 2.0204,
      "step": 330000
    },
    {
      "epoch": 3.6,
      "eval_accuracy": 0.623003453753276,
      "eval_loss": 1.9287109375,
      "eval_runtime": 1034.59,
      "eval_samples_per_second": 298.047,
      "eval_steps_per_second": 4.658,
      "step": 330000
    },
    {
      "epoch": 3.61,
      "learning_rate": 1.4281358885017423e-05,
      "loss": 2.0244,
      "step": 330500
    },
    {
      "epoch": 3.62,
      "learning_rate": 1.4225743768426697e-05,
      "loss": 2.0269,
      "step": 331000
    },
    {
      "epoch": 3.62,
      "learning_rate": 1.4169905297954078e-05,
      "loss": 2.0196,
      "step": 331500
    },
    {
      "epoch": 3.63,
      "learning_rate": 1.4114290181363351e-05,
      "loss": 2.0203,
      "step": 332000
    },
    {
      "epoch": 3.63,
      "learning_rate": 1.4058451710890736e-05,
      "loss": 2.0221,
      "step": 332500
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.400283659430001e-05,
      "loss": 2.0196,
      "step": 333000
    },
    {
      "epoch": 3.64,
      "learning_rate": 1.3946998123827392e-05,
      "loss": 2.029,
      "step": 333500
    },
    {
      "epoch": 3.65,
      "learning_rate": 1.3891383007236666e-05,
      "loss": 2.0204,
      "step": 334000
    },
    {
      "epoch": 3.65,
      "learning_rate": 1.383554453676405e-05,
      "loss": 2.0221,
      "step": 334500
    },
    {
      "epoch": 3.66,
      "learning_rate": 1.3779929420173324e-05,
      "loss": 2.0165,
      "step": 335000
    },
    {
      "epoch": 3.66,
      "learning_rate": 1.3724090949700707e-05,
      "loss": 2.0222,
      "step": 335500
    },
    {
      "epoch": 3.67,
      "learning_rate": 1.3668475833109982e-05,
      "loss": 2.0171,
      "step": 336000
    },
    {
      "epoch": 3.68,
      "learning_rate": 1.3612637362637362e-05,
      "loss": 2.0201,
      "step": 336500
    },
    {
      "epoch": 3.68,
      "learning_rate": 1.3557022246046636e-05,
      "loss": 2.0166,
      "step": 337000
    },
    {
      "epoch": 3.69,
      "learning_rate": 1.350118377557402e-05,
      "loss": 2.0137,
      "step": 337500
    },
    {
      "epoch": 3.69,
      "learning_rate": 1.3445568658983293e-05,
      "loss": 2.0142,
      "step": 338000
    },
    {
      "epoch": 3.7,
      "learning_rate": 1.3389730188510678e-05,
      "loss": 2.0155,
      "step": 338500
    },
    {
      "epoch": 3.7,
      "learning_rate": 1.333411507191995e-05,
      "loss": 2.0177,
      "step": 339000
    },
    {
      "epoch": 3.71,
      "learning_rate": 1.3278276601447334e-05,
      "loss": 2.0164,
      "step": 339500
    },
    {
      "epoch": 3.71,
      "learning_rate": 1.3222661484856607e-05,
      "loss": 2.0142,
      "step": 340000
    },
    {
      "epoch": 3.71,
      "eval_accuracy": 0.6242913468679321,
      "eval_loss": 1.919921875,
      "eval_runtime": 1063.0076,
      "eval_samples_per_second": 290.079,
      "eval_steps_per_second": 4.533,
      "step": 340000
    },
    {
      "epoch": 3.72,
      "learning_rate": 1.3166823014383992e-05,
      "loss": 2.0135,
      "step": 340500
    },
    {
      "epoch": 3.72,
      "learning_rate": 1.3111207897793265e-05,
      "loss": 2.018,
      "step": 341000
    },
    {
      "epoch": 3.73,
      "learning_rate": 1.3055369427320647e-05,
      "loss": 2.012,
      "step": 341500
    },
    {
      "epoch": 3.74,
      "learning_rate": 1.2999754310729923e-05,
      "loss": 2.0122,
      "step": 342000
    },
    {
      "epoch": 3.74,
      "learning_rate": 1.2943915840257303e-05,
      "loss": 2.0152,
      "step": 342500
    },
    {
      "epoch": 3.75,
      "learning_rate": 1.2888300723666578e-05,
      "loss": 2.0103,
      "step": 343000
    },
    {
      "epoch": 3.75,
      "learning_rate": 1.2832462253193961e-05,
      "loss": 2.0105,
      "step": 343500
    },
    {
      "epoch": 3.76,
      "learning_rate": 1.2776847136603234e-05,
      "loss": 2.0157,
      "step": 344000
    },
    {
      "epoch": 3.76,
      "learning_rate": 1.2721008666130619e-05,
      "loss": 2.0107,
      "step": 344500
    },
    {
      "epoch": 3.77,
      "learning_rate": 1.2665393549539892e-05,
      "loss": 2.0136,
      "step": 345000
    },
    {
      "epoch": 3.77,
      "learning_rate": 1.2609555079067275e-05,
      "loss": 2.008,
      "step": 345500
    },
    {
      "epoch": 3.78,
      "learning_rate": 1.2553939962476548e-05,
      "loss": 2.0046,
      "step": 346000
    },
    {
      "epoch": 3.79,
      "learning_rate": 1.2498101492003932e-05,
      "loss": 2.0123,
      "step": 346500
    },
    {
      "epoch": 3.79,
      "learning_rate": 1.2442486375413205e-05,
      "loss": 2.0093,
      "step": 347000
    },
    {
      "epoch": 3.8,
      "learning_rate": 1.238664790494059e-05,
      "loss": 2.0109,
      "step": 347500
    },
    {
      "epoch": 3.8,
      "learning_rate": 1.2331032788349863e-05,
      "loss": 2.0073,
      "step": 348000
    },
    {
      "epoch": 3.81,
      "learning_rate": 1.2275194317877244e-05,
      "loss": 2.0074,
      "step": 348500
    },
    {
      "epoch": 3.81,
      "learning_rate": 1.221957920128652e-05,
      "loss": 2.0117,
      "step": 349000
    },
    {
      "epoch": 3.82,
      "learning_rate": 1.2163740730813902e-05,
      "loss": 2.0087,
      "step": 349500
    },
    {
      "epoch": 3.82,
      "learning_rate": 1.2108125614223176e-05,
      "loss": 2.0021,
      "step": 350000
    },
    {
      "epoch": 3.82,
      "eval_accuracy": 0.6256724294842592,
      "eval_loss": 1.912109375,
      "eval_runtime": 1058.3425,
      "eval_samples_per_second": 291.357,
      "eval_steps_per_second": 4.553,
      "step": 350000
    },
    {
      "epoch": 3.83,
      "learning_rate": 1.2052287143750559e-05,
      "loss": 2.0188,
      "step": 350500
    },
    {
      "epoch": 3.83,
      "learning_rate": 1.1996672027159834e-05,
      "loss": 2.0063,
      "step": 351000
    },
    {
      "epoch": 3.84,
      "learning_rate": 1.1940833556687215e-05,
      "loss": 2.0098,
      "step": 351500
    },
    {
      "epoch": 3.85,
      "learning_rate": 1.188521844009649e-05,
      "loss": 2.003,
      "step": 352000
    },
    {
      "epoch": 3.85,
      "learning_rate": 1.1829379969623873e-05,
      "loss": 2.0034,
      "step": 352500
    },
    {
      "epoch": 3.86,
      "learning_rate": 1.1773764853033146e-05,
      "loss": 2.006,
      "step": 353000
    },
    {
      "epoch": 3.86,
      "learning_rate": 1.171792638256053e-05,
      "loss": 2.0067,
      "step": 353500
    },
    {
      "epoch": 3.87,
      "learning_rate": 1.1662311265969803e-05,
      "loss": 2.0051,
      "step": 354000
    },
    {
      "epoch": 3.87,
      "learning_rate": 1.1606472795497186e-05,
      "loss": 2.0044,
      "step": 354500
    },
    {
      "epoch": 3.88,
      "learning_rate": 1.155085767890646e-05,
      "loss": 2.0028,
      "step": 355000
    },
    {
      "epoch": 3.88,
      "learning_rate": 1.1495019208433844e-05,
      "loss": 2.0097,
      "step": 355500
    },
    {
      "epoch": 3.89,
      "learning_rate": 1.1439404091843117e-05,
      "loss": 2.0048,
      "step": 356000
    },
    {
      "epoch": 3.89,
      "learning_rate": 1.13835656213705e-05,
      "loss": 2.0042,
      "step": 356500
    },
    {
      "epoch": 3.9,
      "learning_rate": 1.1327950504779773e-05,
      "loss": 2.0022,
      "step": 357000
    },
    {
      "epoch": 3.91,
      "learning_rate": 1.1272112034307156e-05,
      "loss": 2.0073,
      "step": 357500
    },
    {
      "epoch": 3.91,
      "learning_rate": 1.1216496917716431e-05,
      "loss": 1.9993,
      "step": 358000
    },
    {
      "epoch": 3.92,
      "learning_rate": 1.1160658447243813e-05,
      "loss": 1.9994,
      "step": 358500
    },
    {
      "epoch": 3.92,
      "learning_rate": 1.1105043330653088e-05,
      "loss": 1.9986,
      "step": 359000
    },
    {
      "epoch": 3.93,
      "learning_rate": 1.104920486018047e-05,
      "loss": 2.0043,
      "step": 359500
    },
    {
      "epoch": 3.93,
      "learning_rate": 1.0993589743589744e-05,
      "loss": 2.006,
      "step": 360000
    },
    {
      "epoch": 3.93,
      "eval_accuracy": 0.6264209623539464,
      "eval_loss": 1.904296875,
      "eval_runtime": 1061.3214,
      "eval_samples_per_second": 290.54,
      "eval_steps_per_second": 4.541,
      "step": 360000
    },
    {
      "epoch": 3.94,
      "learning_rate": 1.0937751273117127e-05,
      "loss": 2.0034,
      "step": 360500
    },
    {
      "epoch": 3.94,
      "learning_rate": 1.0882136156526402e-05,
      "loss": 2.0003,
      "step": 361000
    },
    {
      "epoch": 3.95,
      "learning_rate": 1.0826297686053783e-05,
      "loss": 2.0065,
      "step": 361500
    },
    {
      "epoch": 3.95,
      "learning_rate": 1.0770682569463057e-05,
      "loss": 2.0013,
      "step": 362000
    },
    {
      "epoch": 3.96,
      "learning_rate": 1.0714844098990442e-05,
      "loss": 1.9985,
      "step": 362500
    },
    {
      "epoch": 3.97,
      "learning_rate": 1.0659228982399715e-05,
      "loss": 2.0011,
      "step": 363000
    },
    {
      "epoch": 3.97,
      "learning_rate": 1.0603390511927098e-05,
      "loss": 1.9928,
      "step": 363500
    },
    {
      "epoch": 3.98,
      "learning_rate": 1.0547775395336373e-05,
      "loss": 1.9994,
      "step": 364000
    },
    {
      "epoch": 3.98,
      "learning_rate": 1.0491936924863754e-05,
      "loss": 2.0,
      "step": 364500
    },
    {
      "epoch": 3.99,
      "learning_rate": 1.0436321808273027e-05,
      "loss": 1.9952,
      "step": 365000
    },
    {
      "epoch": 3.99,
      "learning_rate": 1.0380483337800412e-05,
      "loss": 2.0017,
      "step": 365500
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.0324868221209685e-05,
      "loss": 1.9979,
      "step": 366000
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.0269029750737069e-05,
      "loss": 1.9937,
      "step": 366500
    },
    {
      "epoch": 4.01,
      "learning_rate": 1.0213414634146342e-05,
      "loss": 1.993,
      "step": 367000
    },
    {
      "epoch": 4.01,
      "learning_rate": 1.0157576163673725e-05,
      "loss": 1.9951,
      "step": 367500
    },
    {
      "epoch": 4.02,
      "learning_rate": 1.0101961047082998e-05,
      "loss": 1.993,
      "step": 368000
    },
    {
      "epoch": 4.03,
      "learning_rate": 1.0046122576610383e-05,
      "loss": 1.9916,
      "step": 368500
    },
    {
      "epoch": 4.03,
      "learning_rate": 9.990507460019656e-06,
      "loss": 1.997,
      "step": 369000
    },
    {
      "epoch": 4.04,
      "learning_rate": 9.93466898954704e-06,
      "loss": 1.9905,
      "step": 369500
    },
    {
      "epoch": 4.04,
      "learning_rate": 9.879053872956312e-06,
      "loss": 1.9917,
      "step": 370000
    },
    {
      "epoch": 4.04,
      "eval_accuracy": 0.6274146236467327,
      "eval_loss": 1.8984375,
      "eval_runtime": 1044.7732,
      "eval_samples_per_second": 295.142,
      "eval_steps_per_second": 4.612,
      "step": 370000
    },
    {
      "epoch": 4.05,
      "learning_rate": 9.823215402483696e-06,
      "loss": 1.9881,
      "step": 370500
    },
    {
      "epoch": 4.05,
      "learning_rate": 9.767600285892969e-06,
      "loss": 1.9896,
      "step": 371000
    },
    {
      "epoch": 4.06,
      "learning_rate": 9.711761815420352e-06,
      "loss": 1.9941,
      "step": 371500
    },
    {
      "epoch": 4.06,
      "learning_rate": 9.656146698829627e-06,
      "loss": 1.991,
      "step": 372000
    },
    {
      "epoch": 4.07,
      "learning_rate": 9.60030822835701e-06,
      "loss": 1.9904,
      "step": 372500
    },
    {
      "epoch": 4.07,
      "learning_rate": 9.544693111766283e-06,
      "loss": 1.992,
      "step": 373000
    },
    {
      "epoch": 4.08,
      "learning_rate": 9.488854641293666e-06,
      "loss": 1.9917,
      "step": 373500
    },
    {
      "epoch": 4.09,
      "learning_rate": 9.43323952470294e-06,
      "loss": 1.9916,
      "step": 374000
    },
    {
      "epoch": 4.09,
      "learning_rate": 9.377401054230323e-06,
      "loss": 1.9854,
      "step": 374500
    },
    {
      "epoch": 4.1,
      "learning_rate": 9.321785937639596e-06,
      "loss": 1.9907,
      "step": 375000
    },
    {
      "epoch": 4.1,
      "learning_rate": 9.26594746716698e-06,
      "loss": 1.9885,
      "step": 375500
    },
    {
      "epoch": 4.11,
      "learning_rate": 9.210332350576254e-06,
      "loss": 1.9847,
      "step": 376000
    },
    {
      "epoch": 4.11,
      "learning_rate": 9.154493880103637e-06,
      "loss": 1.9888,
      "step": 376500
    },
    {
      "epoch": 4.12,
      "learning_rate": 9.09887876351291e-06,
      "loss": 1.988,
      "step": 377000
    },
    {
      "epoch": 4.12,
      "learning_rate": 9.043040293040293e-06,
      "loss": 1.9891,
      "step": 377500
    },
    {
      "epoch": 4.13,
      "learning_rate": 8.987425176449567e-06,
      "loss": 1.9868,
      "step": 378000
    },
    {
      "epoch": 4.13,
      "learning_rate": 8.931586705976951e-06,
      "loss": 1.9925,
      "step": 378500
    },
    {
      "epoch": 4.14,
      "learning_rate": 8.875971589386225e-06,
      "loss": 1.9871,
      "step": 379000
    },
    {
      "epoch": 4.15,
      "learning_rate": 8.820133118913606e-06,
      "loss": 1.986,
      "step": 379500
    },
    {
      "epoch": 4.15,
      "learning_rate": 8.764518002322881e-06,
      "loss": 1.9881,
      "step": 380000
    },
    {
      "epoch": 4.15,
      "eval_accuracy": 0.6284240406230363,
      "eval_loss": 1.8916015625,
      "eval_runtime": 1042.4014,
      "eval_samples_per_second": 295.813,
      "eval_steps_per_second": 4.623,
      "step": 380000
    },
    {
      "epoch": 4.16,
      "learning_rate": 8.708679531850264e-06,
      "loss": 1.9818,
      "step": 380500
    },
    {
      "epoch": 4.16,
      "learning_rate": 8.653064415259537e-06,
      "loss": 1.9863,
      "step": 381000
    },
    {
      "epoch": 4.17,
      "learning_rate": 8.597225944786922e-06,
      "loss": 1.9852,
      "step": 381500
    },
    {
      "epoch": 4.17,
      "learning_rate": 8.541610828196195e-06,
      "loss": 1.9832,
      "step": 382000
    },
    {
      "epoch": 4.18,
      "learning_rate": 8.485772357723577e-06,
      "loss": 1.9852,
      "step": 382500
    },
    {
      "epoch": 4.18,
      "learning_rate": 8.430157241132852e-06,
      "loss": 1.9855,
      "step": 383000
    },
    {
      "epoch": 4.19,
      "learning_rate": 8.374318770660235e-06,
      "loss": 1.9854,
      "step": 383500
    },
    {
      "epoch": 4.19,
      "learning_rate": 8.318703654069508e-06,
      "loss": 1.9792,
      "step": 384000
    },
    {
      "epoch": 4.2,
      "learning_rate": 8.262865183596891e-06,
      "loss": 1.9818,
      "step": 384500
    },
    {
      "epoch": 4.21,
      "learning_rate": 8.207250067006166e-06,
      "loss": 1.9831,
      "step": 385000
    },
    {
      "epoch": 4.21,
      "learning_rate": 8.151411596533547e-06,
      "loss": 1.9811,
      "step": 385500
    },
    {
      "epoch": 4.22,
      "learning_rate": 8.095796479942822e-06,
      "loss": 1.9786,
      "step": 386000
    },
    {
      "epoch": 4.22,
      "learning_rate": 8.039958009470205e-06,
      "loss": 1.9884,
      "step": 386500
    },
    {
      "epoch": 4.23,
      "learning_rate": 7.984342892879479e-06,
      "loss": 1.9838,
      "step": 387000
    },
    {
      "epoch": 4.23,
      "learning_rate": 7.928504422406862e-06,
      "loss": 1.9821,
      "step": 387500
    },
    {
      "epoch": 4.24,
      "learning_rate": 7.872889305816135e-06,
      "loss": 1.9871,
      "step": 388000
    },
    {
      "epoch": 4.24,
      "learning_rate": 7.817050835343518e-06,
      "loss": 1.9848,
      "step": 388500
    },
    {
      "epoch": 4.25,
      "learning_rate": 7.761435718752793e-06,
      "loss": 1.9854,
      "step": 389000
    },
    {
      "epoch": 4.25,
      "learning_rate": 7.705597248280176e-06,
      "loss": 1.9811,
      "step": 389500
    },
    {
      "epoch": 4.26,
      "learning_rate": 7.64998213168945e-06,
      "loss": 1.9843,
      "step": 390000
    },
    {
      "epoch": 4.26,
      "eval_accuracy": 0.6291420100522126,
      "eval_loss": 1.88671875,
      "eval_runtime": 1042.3765,
      "eval_samples_per_second": 295.82,
      "eval_steps_per_second": 4.623,
      "step": 390000
    },
    {
      "epoch": 4.27,
      "learning_rate": 7.594143661216832e-06,
      "loss": 1.9827,
      "step": 390500
    },
    {
      "epoch": 4.27,
      "learning_rate": 7.538528544626106e-06,
      "loss": 1.9825,
      "step": 391000
    },
    {
      "epoch": 4.28,
      "learning_rate": 7.48269007415349e-06,
      "loss": 1.9772,
      "step": 391500
    },
    {
      "epoch": 4.28,
      "learning_rate": 7.427074957562763e-06,
      "loss": 1.9788,
      "step": 392000
    },
    {
      "epoch": 4.29,
      "learning_rate": 7.371236487090145e-06,
      "loss": 1.9803,
      "step": 392500
    },
    {
      "epoch": 4.29,
      "learning_rate": 7.31562137049942e-06,
      "loss": 1.9762,
      "step": 393000
    },
    {
      "epoch": 4.3,
      "learning_rate": 7.259782900026802e-06,
      "loss": 1.9797,
      "step": 393500
    },
    {
      "epoch": 4.3,
      "learning_rate": 7.204167783436076e-06,
      "loss": 1.98,
      "step": 394000
    },
    {
      "epoch": 4.31,
      "learning_rate": 7.1483293129634604e-06,
      "loss": 1.9782,
      "step": 394500
    },
    {
      "epoch": 4.31,
      "learning_rate": 7.092714196372734e-06,
      "loss": 1.9711,
      "step": 395000
    },
    {
      "epoch": 4.32,
      "learning_rate": 7.036875725900116e-06,
      "loss": 1.9765,
      "step": 395500
    },
    {
      "epoch": 4.33,
      "learning_rate": 6.98126060930939e-06,
      "loss": 1.9789,
      "step": 396000
    },
    {
      "epoch": 4.33,
      "learning_rate": 6.925422138836773e-06,
      "loss": 1.9769,
      "step": 396500
    },
    {
      "epoch": 4.34,
      "learning_rate": 6.869807022246047e-06,
      "loss": 1.9765,
      "step": 397000
    },
    {
      "epoch": 4.34,
      "learning_rate": 6.8139685517734294e-06,
      "loss": 1.9808,
      "step": 397500
    },
    {
      "epoch": 4.35,
      "learning_rate": 6.758353435182704e-06,
      "loss": 1.9831,
      "step": 398000
    },
    {
      "epoch": 4.35,
      "learning_rate": 6.702514964710087e-06,
      "loss": 1.9783,
      "step": 398500
    },
    {
      "epoch": 4.36,
      "learning_rate": 6.646899848119361e-06,
      "loss": 1.9754,
      "step": 399000
    },
    {
      "epoch": 4.36,
      "learning_rate": 6.591061377646744e-06,
      "loss": 1.973,
      "step": 399500
    },
    {
      "epoch": 4.37,
      "learning_rate": 6.535446261056018e-06,
      "loss": 1.977,
      "step": 400000
    },
    {
      "epoch": 4.37,
      "eval_accuracy": 0.6300586017850129,
      "eval_loss": 1.880859375,
      "eval_runtime": 1042.6143,
      "eval_samples_per_second": 295.753,
      "eval_steps_per_second": 4.622,
      "step": 400000
    },
    {
      "epoch": 4.37,
      "learning_rate": 6.4796077905834e-06,
      "loss": 1.9774,
      "step": 400500
    },
    {
      "epoch": 4.38,
      "learning_rate": 6.423992673992673e-06,
      "loss": 1.9796,
      "step": 401000
    },
    {
      "epoch": 4.39,
      "learning_rate": 6.368154203520057e-06,
      "loss": 1.976,
      "step": 401500
    },
    {
      "epoch": 4.39,
      "learning_rate": 6.312539086929331e-06,
      "loss": 1.9758,
      "step": 402000
    },
    {
      "epoch": 4.4,
      "learning_rate": 6.2567006164567145e-06,
      "loss": 1.9818,
      "step": 402500
    },
    {
      "epoch": 4.4,
      "learning_rate": 6.201085499865988e-06,
      "loss": 1.9732,
      "step": 403000
    },
    {
      "epoch": 4.41,
      "learning_rate": 6.145247029393371e-06,
      "loss": 1.9763,
      "step": 403500
    },
    {
      "epoch": 4.41,
      "learning_rate": 6.089631912802645e-06,
      "loss": 1.9756,
      "step": 404000
    },
    {
      "epoch": 4.42,
      "learning_rate": 6.033793442330028e-06,
      "loss": 1.9711,
      "step": 404500
    },
    {
      "epoch": 4.42,
      "learning_rate": 5.978178325739302e-06,
      "loss": 1.9705,
      "step": 405000
    },
    {
      "epoch": 4.43,
      "learning_rate": 5.922339855266685e-06,
      "loss": 1.9762,
      "step": 405500
    },
    {
      "epoch": 4.44,
      "learning_rate": 5.866724738675958e-06,
      "loss": 1.9721,
      "step": 406000
    },
    {
      "epoch": 4.44,
      "learning_rate": 5.8108862682033416e-06,
      "loss": 1.9738,
      "step": 406500
    },
    {
      "epoch": 4.45,
      "learning_rate": 5.755271151612615e-06,
      "loss": 1.9709,
      "step": 407000
    },
    {
      "epoch": 4.45,
      "learning_rate": 5.699432681139998e-06,
      "loss": 1.967,
      "step": 407500
    },
    {
      "epoch": 4.46,
      "learning_rate": 5.643817564549273e-06,
      "loss": 1.9784,
      "step": 408000
    },
    {
      "epoch": 4.46,
      "learning_rate": 5.587979094076656e-06,
      "loss": 1.9764,
      "step": 408500
    },
    {
      "epoch": 4.47,
      "learning_rate": 5.532363977485929e-06,
      "loss": 1.9702,
      "step": 409000
    },
    {
      "epoch": 4.47,
      "learning_rate": 5.476525507013312e-06,
      "loss": 1.9732,
      "step": 409500
    },
    {
      "epoch": 4.48,
      "learning_rate": 5.4209103904225854e-06,
      "loss": 1.9697,
      "step": 410000
    },
    {
      "epoch": 4.48,
      "eval_accuracy": 0.6306101079076281,
      "eval_loss": 1.876953125,
      "eval_runtime": 1037.0911,
      "eval_samples_per_second": 297.328,
      "eval_steps_per_second": 4.647,
      "step": 410000
    },
    {
      "epoch": 4.48,
      "learning_rate": 5.365071919949969e-06,
      "loss": 1.976,
      "step": 410500
    },
    {
      "epoch": 4.49,
      "learning_rate": 5.309456803359243e-06,
      "loss": 1.9726,
      "step": 411000
    },
    {
      "epoch": 4.5,
      "learning_rate": 5.253618332886626e-06,
      "loss": 1.9745,
      "step": 411500
    },
    {
      "epoch": 4.5,
      "learning_rate": 5.198003216295899e-06,
      "loss": 1.9744,
      "step": 412000
    },
    {
      "epoch": 4.51,
      "learning_rate": 5.142164745823282e-06,
      "loss": 1.975,
      "step": 412500
    },
    {
      "epoch": 4.51,
      "learning_rate": 5.086549629232556e-06,
      "loss": 1.9705,
      "step": 413000
    },
    {
      "epoch": 4.52,
      "learning_rate": 5.030711158759939e-06,
      "loss": 1.9658,
      "step": 413500
    },
    {
      "epoch": 4.52,
      "learning_rate": 4.975096042169213e-06,
      "loss": 1.9701,
      "step": 414000
    },
    {
      "epoch": 4.53,
      "learning_rate": 4.9192575716965965e-06,
      "loss": 1.9623,
      "step": 414500
    },
    {
      "epoch": 4.53,
      "learning_rate": 4.86364245510587e-06,
      "loss": 1.9731,
      "step": 415000
    },
    {
      "epoch": 4.54,
      "learning_rate": 4.807803984633253e-06,
      "loss": 1.9727,
      "step": 415500
    },
    {
      "epoch": 4.54,
      "learning_rate": 4.752188868042527e-06,
      "loss": 1.9624,
      "step": 416000
    },
    {
      "epoch": 4.55,
      "learning_rate": 4.69635039756991e-06,
      "loss": 1.9669,
      "step": 416500
    },
    {
      "epoch": 4.56,
      "learning_rate": 4.640735280979184e-06,
      "loss": 1.9641,
      "step": 417000
    },
    {
      "epoch": 4.56,
      "learning_rate": 4.584896810506567e-06,
      "loss": 1.9664,
      "step": 417500
    },
    {
      "epoch": 4.57,
      "learning_rate": 4.52928169391584e-06,
      "loss": 1.9722,
      "step": 418000
    },
    {
      "epoch": 4.57,
      "learning_rate": 4.4734432234432235e-06,
      "loss": 1.9646,
      "step": 418500
    },
    {
      "epoch": 4.58,
      "learning_rate": 4.4178281068524975e-06,
      "loss": 1.971,
      "step": 419000
    },
    {
      "epoch": 4.58,
      "learning_rate": 4.361989636379881e-06,
      "loss": 1.9666,
      "step": 419500
    },
    {
      "epoch": 4.59,
      "learning_rate": 4.306374519789154e-06,
      "loss": 1.9655,
      "step": 420000
    },
    {
      "epoch": 4.59,
      "eval_accuracy": 0.6312552391745295,
      "eval_loss": 1.8740234375,
      "eval_runtime": 1117.5437,
      "eval_samples_per_second": 275.923,
      "eval_steps_per_second": 4.312,
      "step": 420000
    },
    {
      "epoch": 4.59,
      "learning_rate": 4.250536049316537e-06,
      "loss": 1.9681,
      "step": 420500
    },
    {
      "epoch": 4.6,
      "learning_rate": 4.194920932725811e-06,
      "loss": 1.969,
      "step": 421000
    },
    {
      "epoch": 4.6,
      "learning_rate": 4.139082462253194e-06,
      "loss": 1.9663,
      "step": 421500
    },
    {
      "epoch": 4.61,
      "learning_rate": 4.083467345662468e-06,
      "loss": 1.9684,
      "step": 422000
    },
    {
      "epoch": 4.62,
      "learning_rate": 4.0276288751898514e-06,
      "loss": 1.9675,
      "step": 422500
    },
    {
      "epoch": 4.62,
      "learning_rate": 3.972013758599125e-06,
      "loss": 1.9649,
      "step": 423000
    },
    {
      "epoch": 4.63,
      "learning_rate": 3.916175288126508e-06,
      "loss": 1.9665,
      "step": 423500
    },
    {
      "epoch": 4.63,
      "learning_rate": 3.860560171535781e-06,
      "loss": 1.9664,
      "step": 424000
    },
    {
      "epoch": 4.64,
      "learning_rate": 3.8047217010631645e-06,
      "loss": 1.9637,
      "step": 424500
    },
    {
      "epoch": 4.64,
      "learning_rate": 3.7491065844724385e-06,
      "loss": 1.963,
      "step": 425000
    },
    {
      "epoch": 4.65,
      "learning_rate": 3.6932681139998217e-06,
      "loss": 1.9623,
      "step": 425500
    },
    {
      "epoch": 4.65,
      "learning_rate": 3.6376529974090953e-06,
      "loss": 1.9651,
      "step": 426000
    },
    {
      "epoch": 4.66,
      "learning_rate": 3.5818145269364785e-06,
      "loss": 1.9647,
      "step": 426500
    },
    {
      "epoch": 4.66,
      "learning_rate": 3.526199410345752e-06,
      "loss": 1.9617,
      "step": 427000
    },
    {
      "epoch": 4.67,
      "learning_rate": 3.4703609398731352e-06,
      "loss": 1.9672,
      "step": 427500
    },
    {
      "epoch": 4.68,
      "learning_rate": 3.414745823282409e-06,
      "loss": 1.9645,
      "step": 428000
    },
    {
      "epoch": 4.68,
      "learning_rate": 3.358907352809792e-06,
      "loss": 1.9624,
      "step": 428500
    },
    {
      "epoch": 4.69,
      "learning_rate": 3.303292236219065e-06,
      "loss": 1.9681,
      "step": 429000
    },
    {
      "epoch": 4.69,
      "learning_rate": 3.2474537657464483e-06,
      "loss": 1.9648,
      "step": 429500
    },
    {
      "epoch": 4.7,
      "learning_rate": 3.1918386491557228e-06,
      "loss": 1.9649,
      "step": 430000
    },
    {
      "epoch": 4.7,
      "eval_accuracy": 0.6319568161801343,
      "eval_loss": 1.869140625,
      "eval_runtime": 1118.6445,
      "eval_samples_per_second": 275.651,
      "eval_steps_per_second": 4.308,
      "step": 430000
    },
    {
      "epoch": 4.7,
      "learning_rate": 3.136000178683106e-06,
      "loss": 1.9599,
      "step": 430500
    },
    {
      "epoch": 4.71,
      "learning_rate": 3.0803850620923795e-06,
      "loss": 1.9689,
      "step": 431000
    },
    {
      "epoch": 4.71,
      "learning_rate": 3.0245465916197627e-06,
      "loss": 1.9665,
      "step": 431500
    },
    {
      "epoch": 4.72,
      "learning_rate": 2.968931475029036e-06,
      "loss": 1.9623,
      "step": 432000
    },
    {
      "epoch": 4.72,
      "learning_rate": 2.913093004556419e-06,
      "loss": 1.9615,
      "step": 432500
    },
    {
      "epoch": 4.73,
      "learning_rate": 2.857477887965693e-06,
      "loss": 1.9636,
      "step": 433000
    },
    {
      "epoch": 4.74,
      "learning_rate": 2.8016394174930762e-06,
      "loss": 1.956,
      "step": 433500
    },
    {
      "epoch": 4.74,
      "learning_rate": 2.74602430090235e-06,
      "loss": 1.9613,
      "step": 434000
    },
    {
      "epoch": 4.75,
      "learning_rate": 2.690185830429733e-06,
      "loss": 1.9601,
      "step": 434500
    },
    {
      "epoch": 4.75,
      "learning_rate": 2.6345707138390066e-06,
      "loss": 1.9599,
      "step": 435000
    },
    {
      "epoch": 4.76,
      "learning_rate": 2.5787322433663898e-06,
      "loss": 1.9652,
      "step": 435500
    },
    {
      "epoch": 4.76,
      "learning_rate": 2.5231171267756633e-06,
      "loss": 1.9626,
      "step": 436000
    },
    {
      "epoch": 4.77,
      "learning_rate": 2.4672786563030465e-06,
      "loss": 1.9586,
      "step": 436500
    },
    {
      "epoch": 4.77,
      "learning_rate": 2.4116635397123205e-06,
      "loss": 1.9618,
      "step": 437000
    },
    {
      "epoch": 4.78,
      "learning_rate": 2.3558250692397037e-06,
      "loss": 1.9611,
      "step": 437500
    },
    {
      "epoch": 4.78,
      "learning_rate": 2.3002099526489773e-06,
      "loss": 1.9603,
      "step": 438000
    },
    {
      "epoch": 4.79,
      "learning_rate": 2.2443714821763605e-06,
      "loss": 1.9565,
      "step": 438500
    },
    {
      "epoch": 4.8,
      "learning_rate": 2.188756365585634e-06,
      "loss": 1.9609,
      "step": 439000
    },
    {
      "epoch": 4.8,
      "learning_rate": 2.1329178951130172e-06,
      "loss": 1.9569,
      "step": 439500
    },
    {
      "epoch": 4.81,
      "learning_rate": 2.077302778522291e-06,
      "loss": 1.9622,
      "step": 440000
    },
    {
      "epoch": 4.81,
      "eval_accuracy": 0.6323656721212422,
      "eval_loss": 1.8662109375,
      "eval_runtime": 1118.8914,
      "eval_samples_per_second": 275.591,
      "eval_steps_per_second": 4.307,
      "step": 440000
    },
    {
      "epoch": 4.81,
      "learning_rate": 2.021464308049674e-06,
      "loss": 1.9538,
      "step": 440500
    },
    {
      "epoch": 4.82,
      "learning_rate": 1.965849191458948e-06,
      "loss": 1.9593,
      "step": 441000
    },
    {
      "epoch": 4.82,
      "learning_rate": 1.910010720986331e-06,
      "loss": 1.9642,
      "step": 441500
    },
    {
      "epoch": 4.83,
      "learning_rate": 1.8543956043956045e-06,
      "loss": 1.9562,
      "step": 442000
    },
    {
      "epoch": 4.83,
      "learning_rate": 1.7985571339229877e-06,
      "loss": 1.9582,
      "step": 442500
    },
    {
      "epoch": 4.84,
      "learning_rate": 1.742942017332261e-06,
      "loss": 1.9576,
      "step": 443000
    },
    {
      "epoch": 4.84,
      "learning_rate": 1.6871035468596447e-06,
      "loss": 1.9566,
      "step": 443500
    },
    {
      "epoch": 4.85,
      "learning_rate": 1.6314884302689183e-06,
      "loss": 1.96,
      "step": 444000
    },
    {
      "epoch": 4.86,
      "learning_rate": 1.5756499597963012e-06,
      "loss": 1.963,
      "step": 444500
    },
    {
      "epoch": 4.86,
      "learning_rate": 1.520034843205575e-06,
      "loss": 1.959,
      "step": 445000
    },
    {
      "epoch": 4.87,
      "learning_rate": 1.4641963727329582e-06,
      "loss": 1.9565,
      "step": 445500
    },
    {
      "epoch": 4.87,
      "learning_rate": 1.4085812561422318e-06,
      "loss": 1.9556,
      "step": 446000
    },
    {
      "epoch": 4.88,
      "learning_rate": 1.352742785669615e-06,
      "loss": 1.9551,
      "step": 446500
    },
    {
      "epoch": 4.88,
      "learning_rate": 1.2971276690788888e-06,
      "loss": 1.958,
      "step": 447000
    },
    {
      "epoch": 4.89,
      "learning_rate": 1.241289198606272e-06,
      "loss": 1.9601,
      "step": 447500
    },
    {
      "epoch": 4.89,
      "learning_rate": 1.1856740820155455e-06,
      "loss": 1.9605,
      "step": 448000
    },
    {
      "epoch": 4.9,
      "learning_rate": 1.1298356115429287e-06,
      "loss": 1.9554,
      "step": 448500
    },
    {
      "epoch": 4.9,
      "learning_rate": 1.0742204949522025e-06,
      "loss": 1.9547,
      "step": 449000
    },
    {
      "epoch": 4.91,
      "learning_rate": 1.0183820244795855e-06,
      "loss": 1.9543,
      "step": 449500
    },
    {
      "epoch": 4.92,
      "learning_rate": 9.62766907888859e-07,
      "loss": 1.9539,
      "step": 450000
    },
    {
      "epoch": 4.92,
      "eval_accuracy": 0.6327838084914335,
      "eval_loss": 1.8623046875,
      "eval_runtime": 1037.1505,
      "eval_samples_per_second": 297.311,
      "eval_steps_per_second": 4.646,
      "step": 450000
    },
    {
      "epoch": 4.92,
      "learning_rate": 9.069284374162424e-07,
      "loss": 1.9567,
      "step": 450500
    },
    {
      "epoch": 4.93,
      "learning_rate": 8.513133208255159e-07,
      "loss": 1.9581,
      "step": 451000
    },
    {
      "epoch": 4.93,
      "learning_rate": 7.954748503528992e-07,
      "loss": 1.9539,
      "step": 451500
    },
    {
      "epoch": 4.94,
      "learning_rate": 7.398597337621729e-07,
      "loss": 1.958,
      "step": 452000
    },
    {
      "epoch": 4.94,
      "learning_rate": 6.84021263289556e-07,
      "loss": 1.9592,
      "step": 452500
    },
    {
      "epoch": 4.95,
      "learning_rate": 6.284061466988297e-07,
      "loss": 1.9561,
      "step": 453000
    },
    {
      "epoch": 4.95,
      "learning_rate": 5.725676762262128e-07,
      "loss": 1.9556,
      "step": 453500
    },
    {
      "epoch": 4.96,
      "learning_rate": 5.169525596354864e-07,
      "loss": 1.9576,
      "step": 454000
    },
    {
      "epoch": 4.96,
      "learning_rate": 4.611140891628697e-07,
      "loss": 1.9512,
      "step": 454500
    },
    {
      "epoch": 4.97,
      "learning_rate": 4.054989725721433e-07,
      "loss": 1.9578,
      "step": 455000
    },
    {
      "epoch": 4.98,
      "learning_rate": 3.496605020995265e-07,
      "loss": 1.9538,
      "step": 455500
    },
    {
      "epoch": 4.98,
      "learning_rate": 2.9404538550880015e-07,
      "loss": 1.9564,
      "step": 456000
    },
    {
      "epoch": 4.99,
      "learning_rate": 2.3820691503618333e-07,
      "loss": 1.9589,
      "step": 456500
    },
    {
      "epoch": 4.99,
      "learning_rate": 1.82591798445457e-07,
      "loss": 1.9533,
      "step": 457000
    },
    {
      "epoch": 5.0,
      "learning_rate": 1.2675332797284018e-07,
      "loss": 1.9602,
      "step": 457500
    },
    {
      "epoch": 5.0,
      "step": 457720,
      "total_flos": 7.298372203161059e+18,
      "train_loss": 2.572186891088575,
      "train_runtime": 307115.8978,
      "train_samples_per_second": 95.384,
      "train_steps_per_second": 1.49
    }
  ],
  "max_steps": 457720,
  "num_train_epochs": 5,
  "total_flos": 7.298372203161059e+18,
  "trial_name": null,
  "trial_params": null
}