{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 47.0,
  "eval_steps": 500,
  "global_step": 16029,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.29,
      "learning_rate": 3e-05,
      "loss": 1.574,
      "step": 100
    },
    {
      "epoch": 0.59,
      "learning_rate": 3e-05,
      "loss": 1.474,
      "step": 200
    },
    {
      "epoch": 0.88,
      "learning_rate": 3e-05,
      "loss": 1.4104,
      "step": 300
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.4536923076923077,
      "eval_loss": 3.3574578762054443,
      "eval_runtime": 4.4299,
      "eval_samples_per_second": 112.869,
      "eval_steps_per_second": 14.221,
      "step": 341
    },
    {
      "epoch": 1.0,
      "eval_bleu": 0.08840557310176213,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.058,
      "step": 341
    },
    {
      "epoch": 1.17,
      "learning_rate": 3e-05,
      "loss": 1.3889,
      "step": 400
    },
    {
      "epoch": 1.46,
      "learning_rate": 3e-05,
      "loss": 1.387,
      "step": 500
    },
    {
      "epoch": 1.76,
      "learning_rate": 3e-05,
      "loss": 1.389,
      "step": 600
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.4543589743589744,
      "eval_loss": 3.4179794788360596,
      "eval_runtime": 4.1997,
      "eval_samples_per_second": 119.055,
      "eval_steps_per_second": 15.001,
      "step": 683
    },
    {
      "epoch": 2.0,
      "eval_bleu": 0.09820569565497561,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.056,
      "step": 683
    },
    {
      "epoch": 2.05,
      "learning_rate": 3e-05,
      "loss": 1.3725,
      "step": 700
    },
    {
      "epoch": 2.34,
      "learning_rate": 3e-05,
      "loss": 1.3588,
      "step": 800
    },
    {
      "epoch": 2.64,
      "learning_rate": 3e-05,
      "loss": 1.348,
      "step": 900
    },
    {
      "epoch": 2.93,
      "learning_rate": 3e-05,
      "loss": 1.3414,
      "step": 1000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.45476923076923076,
      "eval_loss": 3.511923313140869,
      "eval_runtime": 4.1985,
      "eval_samples_per_second": 119.089,
      "eval_steps_per_second": 15.005,
      "step": 1024
    },
    {
      "epoch": 3.0,
      "eval_bleu": 0.09889831624589397,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.056,
      "step": 1024
    },
    {
      "epoch": 3.22,
      "learning_rate": 3e-05,
      "loss": 1.3166,
      "step": 1100
    },
    {
      "epoch": 3.51,
      "learning_rate": 3e-05,
      "loss": 1.3035,
      "step": 1200
    },
    {
      "epoch": 3.81,
      "learning_rate": 3e-05,
      "loss": 1.3002,
      "step": 1300
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.45543589743589746,
      "eval_loss": 3.5288071632385254,
      "eval_runtime": 4.4066,
      "eval_samples_per_second": 113.466,
      "eval_steps_per_second": 14.297,
      "step": 1366
    },
    {
      "epoch": 4.0,
      "eval_bleu": 0.11209345982903633,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.054,
      "step": 1366
    },
    {
      "epoch": 4.1,
      "learning_rate": 3e-05,
      "loss": 1.2929,
      "step": 1400
    },
    {
      "epoch": 4.39,
      "learning_rate": 3e-05,
      "loss": 1.2693,
      "step": 1500
    },
    {
      "epoch": 4.69,
      "learning_rate": 3e-05,
      "loss": 1.2697,
      "step": 1600
    },
    {
      "epoch": 4.98,
      "learning_rate": 3e-05,
      "loss": 1.2574,
      "step": 1700
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.45394871794871794,
      "eval_loss": 3.6893365383148193,
      "eval_runtime": 4.1122,
      "eval_samples_per_second": 121.591,
      "eval_steps_per_second": 15.32,
      "step": 1707
    },
    {
      "epoch": 5.0,
      "eval_bleu": 0.12076940663377221,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.056,
      "step": 1707
    },
    {
      "epoch": 5.27,
      "learning_rate": 3e-05,
      "loss": 1.2097,
      "step": 1800
    },
    {
      "epoch": 5.56,
      "learning_rate": 3e-05,
      "loss": 1.2188,
      "step": 1900
    },
    {
      "epoch": 5.86,
      "learning_rate": 3e-05,
      "loss": 1.2258,
      "step": 2000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.45615384615384613,
      "eval_loss": 3.725893497467041,
      "eval_runtime": 4.0989,
      "eval_samples_per_second": 121.984,
      "eval_steps_per_second": 15.37,
      "step": 2049
    },
    {
      "epoch": 6.0,
      "eval_bleu": 0.11884272898137646,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.058,
      "step": 2049
    },
    {
      "epoch": 6.15,
      "learning_rate": 3e-05,
      "loss": 1.205,
      "step": 2100
    },
    {
      "epoch": 6.44,
      "learning_rate": 3e-05,
      "loss": 1.1643,
      "step": 2200
    },
    {
      "epoch": 6.73,
      "learning_rate": 3e-05,
      "loss": 1.1844,
      "step": 2300
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.45594871794871794,
      "eval_loss": 3.724449396133423,
      "eval_runtime": 4.8158,
      "eval_samples_per_second": 103.825,
      "eval_steps_per_second": 13.082,
      "step": 2390
    },
    {
      "epoch": 7.0,
      "eval_bleu": 0.12407896336995003,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.048,
      "step": 2390
    },
    {
      "epoch": 7.03,
      "learning_rate": 3e-05,
      "loss": 1.1771,
      "step": 2400
    },
    {
      "epoch": 7.32,
      "learning_rate": 3e-05,
      "loss": 1.1218,
      "step": 2500
    },
    {
      "epoch": 7.61,
      "learning_rate": 3e-05,
      "loss": 1.1491,
      "step": 2600
    },
    {
      "epoch": 7.91,
      "learning_rate": 3e-05,
      "loss": 1.1363,
      "step": 2700
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.4543589743589744,
      "eval_loss": 3.8138701915740967,
      "eval_runtime": 4.1963,
      "eval_samples_per_second": 119.152,
      "eval_steps_per_second": 15.013,
      "step": 2732
    },
    {
      "epoch": 8.0,
      "eval_bleu": 0.11947629327387163,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.046,
      "step": 2732
    },
    {
      "epoch": 8.2,
      "learning_rate": 3e-05,
      "loss": 1.0858,
      "step": 2800
    },
    {
      "epoch": 8.49,
      "learning_rate": 3e-05,
      "loss": 1.0951,
      "step": 2900
    },
    {
      "epoch": 8.78,
      "learning_rate": 3e-05,
      "loss": 1.0903,
      "step": 3000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.4524102564102564,
      "eval_loss": 3.9115548133850098,
      "eval_runtime": 4.4011,
      "eval_samples_per_second": 113.609,
      "eval_steps_per_second": 14.315,
      "step": 3073
    },
    {
      "epoch": 9.0,
      "eval_bleu": 0.13426062809240774,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.034,
      "step": 3073
    },
    {
      "epoch": 9.08,
      "learning_rate": 3e-05,
      "loss": 1.0822,
      "step": 3100
    },
    {
      "epoch": 9.37,
      "learning_rate": 3e-05,
      "loss": 1.0414,
      "step": 3200
    },
    {
      "epoch": 9.66,
      "learning_rate": 3e-05,
      "loss": 1.0563,
      "step": 3300
    },
    {
      "epoch": 9.96,
      "learning_rate": 3e-05,
      "loss": 1.0538,
      "step": 3400
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.4515897435897436,
      "eval_loss": 3.92203426361084,
      "eval_runtime": 4.4105,
      "eval_samples_per_second": 113.366,
      "eval_steps_per_second": 14.284,
      "step": 3415
    },
    {
      "epoch": 10.0,
      "eval_bleu": 0.11780869016319889,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.032,
      "step": 3415
    },
    {
      "epoch": 10.25,
      "learning_rate": 3e-05,
      "loss": 0.991,
      "step": 3500
    },
    {
      "epoch": 10.54,
      "learning_rate": 3e-05,
      "loss": 1.0019,
      "step": 3600
    },
    {
      "epoch": 10.83,
      "learning_rate": 3e-05,
      "loss": 0.9971,
      "step": 3700
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.45143589743589746,
      "eval_loss": 3.967252492904663,
      "eval_runtime": 4.1003,
      "eval_samples_per_second": 121.943,
      "eval_steps_per_second": 15.365,
      "step": 3756
    },
    {
      "epoch": 11.0,
      "eval_bleu": 0.12407049349871409,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.026,
      "step": 3756
    },
    {
      "epoch": 11.13,
      "learning_rate": 3e-05,
      "loss": 0.9861,
      "step": 3800
    },
    {
      "epoch": 11.42,
      "learning_rate": 3e-05,
      "loss": 0.9436,
      "step": 3900
    },
    {
      "epoch": 11.71,
      "learning_rate": 3e-05,
      "loss": 0.9699,
      "step": 4000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.45076923076923076,
      "eval_loss": 4.033609867095947,
      "eval_runtime": 4.8103,
      "eval_samples_per_second": 103.943,
      "eval_steps_per_second": 13.097,
      "step": 4098
    },
    {
      "epoch": 12.0,
      "eval_bleu": 0.12190601380876163,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.026,
      "step": 4098
    },
    {
      "epoch": 12.01,
      "learning_rate": 3e-05,
      "loss": 0.9783,
      "step": 4100
    },
    {
      "epoch": 12.3,
      "learning_rate": 3e-05,
      "loss": 0.9064,
      "step": 4200
    },
    {
      "epoch": 12.59,
      "learning_rate": 3e-05,
      "loss": 0.9178,
      "step": 4300
    },
    {
      "epoch": 12.88,
      "learning_rate": 3e-05,
      "loss": 0.9235,
      "step": 4400
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.4492820512820513,
      "eval_loss": 4.002023696899414,
      "eval_runtime": 4.8204,
      "eval_samples_per_second": 103.726,
      "eval_steps_per_second": 13.069,
      "step": 4439
    },
    {
      "epoch": 13.0,
      "eval_bleu": 0.13115986163477472,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.026,
      "step": 4439
    },
    {
      "epoch": 13.18,
      "learning_rate": 3e-05,
      "loss": 0.8902,
      "step": 4500
    },
    {
      "epoch": 13.47,
      "learning_rate": 3e-05,
      "loss": 0.8703,
      "step": 4600
    },
    {
      "epoch": 13.76,
      "learning_rate": 3e-05,
      "loss": 0.891,
      "step": 4700
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.44774358974358974,
      "eval_loss": 4.071566104888916,
      "eval_runtime": 4.8316,
      "eval_samples_per_second": 103.485,
      "eval_steps_per_second": 13.039,
      "step": 4781
    },
    {
      "epoch": 14.0,
      "eval_bleu": 0.12164300660650715,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.024,
      "step": 4781
    },
    {
      "epoch": 14.06,
      "learning_rate": 3e-05,
      "loss": 0.8698,
      "step": 4800
    },
    {
      "epoch": 14.35,
      "learning_rate": 3e-05,
      "loss": 0.8338,
      "step": 4900
    },
    {
      "epoch": 14.64,
      "learning_rate": 3e-05,
      "loss": 0.8362,
      "step": 5000
    },
    {
      "epoch": 14.93,
      "learning_rate": 3e-05,
      "loss": 0.845,
      "step": 5100
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.44774358974358974,
      "eval_loss": 4.09920597076416,
      "eval_runtime": 4.4174,
      "eval_samples_per_second": 113.19,
      "eval_steps_per_second": 14.262,
      "step": 5122
    },
    {
      "epoch": 15.0,
      "eval_bleu": 0.1224977607294519,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.02,
      "step": 5122
    },
    {
      "epoch": 15.23,
      "learning_rate": 3e-05,
      "loss": 0.7981,
      "step": 5200
    },
    {
      "epoch": 15.52,
      "learning_rate": 3e-05,
      "loss": 0.7978,
      "step": 5300
    },
    {
      "epoch": 15.81,
      "learning_rate": 3e-05,
      "loss": 0.8009,
      "step": 5400
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.4464102564102564,
      "eval_loss": 4.093270301818848,
      "eval_runtime": 4.102,
      "eval_samples_per_second": 121.892,
      "eval_steps_per_second": 15.358,
      "step": 5464
    },
    {
      "epoch": 16.0,
      "eval_bleu": 0.12201237729153987,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.022,
      "step": 5464
    },
    {
      "epoch": 16.11,
      "learning_rate": 3e-05,
      "loss": 0.7948,
      "step": 5500
    },
    {
      "epoch": 16.4,
      "learning_rate": 3e-05,
      "loss": 0.746,
      "step": 5600
    },
    {
      "epoch": 16.69,
      "learning_rate": 3e-05,
      "loss": 0.7627,
      "step": 5700
    },
    {
      "epoch": 16.98,
      "learning_rate": 3e-05,
      "loss": 0.782,
      "step": 5800
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.44671794871794873,
      "eval_loss": 4.1283488273620605,
      "eval_runtime": 4.84,
      "eval_samples_per_second": 103.305,
      "eval_steps_per_second": 13.016,
      "step": 5805
    },
    {
      "epoch": 17.0,
      "eval_bleu": 0.1263852710208009,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.02,
      "step": 5805
    },
    {
      "epoch": 17.28,
      "learning_rate": 3e-05,
      "loss": 0.7142,
      "step": 5900
    },
    {
      "epoch": 17.57,
      "learning_rate": 3e-05,
      "loss": 0.7253,
      "step": 6000
    },
    {
      "epoch": 17.86,
      "learning_rate": 3e-05,
      "loss": 0.7294,
      "step": 6100
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.44564102564102565,
      "eval_loss": 4.164257049560547,
      "eval_runtime": 4.1037,
      "eval_samples_per_second": 121.84,
      "eval_steps_per_second": 15.352,
      "step": 6147
    },
    {
      "epoch": 18.0,
      "eval_bleu": 0.10715019125675011,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.022,
      "step": 6147
    },
    {
      "epoch": 18.16,
      "learning_rate": 3e-05,
      "loss": 0.7055,
      "step": 6200
    },
    {
      "epoch": 18.45,
      "learning_rate": 3e-05,
      "loss": 0.672,
      "step": 6300
    },
    {
      "epoch": 18.74,
      "learning_rate": 3e-05,
      "loss": 0.6792,
      "step": 6400
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.44487179487179485,
      "eval_loss": 4.185911655426025,
      "eval_runtime": 4.823,
      "eval_samples_per_second": 103.67,
      "eval_steps_per_second": 13.062,
      "step": 6488
    },
    {
      "epoch": 19.0,
      "eval_bleu": 0.11314440439431668,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.02,
      "step": 6488
    },
    {
      "epoch": 19.03,
      "learning_rate": 3e-05,
      "loss": 0.7053,
      "step": 6500
    },
    {
      "epoch": 19.33,
      "learning_rate": 3e-05,
      "loss": 0.6448,
      "step": 6600
    },
    {
      "epoch": 19.62,
      "learning_rate": 3e-05,
      "loss": 0.6448,
      "step": 6700
    },
    {
      "epoch": 19.91,
      "learning_rate": 3e-05,
      "loss": 0.6672,
      "step": 6800
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.4436923076923077,
      "eval_loss": 4.201004981994629,
      "eval_runtime": 4.1985,
      "eval_samples_per_second": 119.09,
      "eval_steps_per_second": 15.005,
      "step": 6830
    },
    {
      "epoch": 20.0,
      "eval_bleu": 0.11913021835274795,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.02,
      "step": 6830
    },
    {
      "epoch": 20.2,
      "learning_rate": 3e-05,
      "loss": 0.6146,
      "step": 6900
    },
    {
      "epoch": 20.5,
      "learning_rate": 3e-05,
      "loss": 0.6102,
      "step": 7000
    },
    {
      "epoch": 20.79,
      "learning_rate": 3e-05,
      "loss": 0.6258,
      "step": 7100
    },
    {
      "epoch": 21.0,
      "eval_accuracy": 0.4428717948717949,
      "eval_loss": 4.230019569396973,
      "eval_runtime": 4.4124,
      "eval_samples_per_second": 113.318,
      "eval_steps_per_second": 14.278,
      "step": 7171
    },
    {
      "epoch": 21.0,
      "eval_bleu": 0.12133891034082767,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.022,
      "step": 7171
    },
    {
      "epoch": 21.08,
      "learning_rate": 3e-05,
      "loss": 0.6149,
      "step": 7200
    },
    {
      "epoch": 21.38,
      "learning_rate": 3e-05,
      "loss": 0.5669,
      "step": 7300
    },
    {
      "epoch": 21.67,
      "learning_rate": 3e-05,
      "loss": 0.5923,
      "step": 7400
    },
    {
      "epoch": 21.96,
      "learning_rate": 3e-05,
      "loss": 0.599,
      "step": 7500
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.4418974358974359,
      "eval_loss": 4.253176689147949,
      "eval_runtime": 4.4121,
      "eval_samples_per_second": 113.326,
      "eval_steps_per_second": 14.279,
      "step": 7513
    },
    {
      "epoch": 22.0,
      "eval_bleu": 0.1179919389999147,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.02,
      "step": 7513
    },
    {
      "epoch": 22.25,
      "learning_rate": 3e-05,
      "loss": 0.536,
      "step": 7600
    },
    {
      "epoch": 22.55,
      "learning_rate": 3e-05,
      "loss": 0.5479,
      "step": 7700
    },
    {
      "epoch": 22.84,
      "learning_rate": 3e-05,
      "loss": 0.5625,
      "step": 7800
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.443025641025641,
      "eval_loss": 4.293737411499023,
      "eval_runtime": 4.4185,
      "eval_samples_per_second": 113.162,
      "eval_steps_per_second": 14.258,
      "step": 7854
    },
    {
      "epoch": 23.0,
      "eval_bleu": 0.10699095482247514,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.022,
      "step": 7854
    },
    {
      "epoch": 23.13,
      "learning_rate": 3e-05,
      "loss": 0.5409,
      "step": 7900
    },
    {
      "epoch": 23.43,
      "learning_rate": 3e-05,
      "loss": 0.5067,
      "step": 8000
    },
    {
      "epoch": 23.72,
      "learning_rate": 3e-05,
      "loss": 0.5267,
      "step": 8100
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.44153846153846155,
      "eval_loss": 4.254815578460693,
      "eval_runtime": 4.2068,
      "eval_samples_per_second": 118.854,
      "eval_steps_per_second": 14.976,
      "step": 8196
    },
    {
      "epoch": 24.0,
      "eval_bleu": 0.10975165733539657,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.024,
      "step": 8196
    },
    {
      "epoch": 24.01,
      "learning_rate": 3e-05,
      "loss": 0.5363,
      "step": 8200
    },
    {
      "epoch": 24.3,
      "learning_rate": 3e-05,
      "loss": 0.4806,
      "step": 8300
    },
    {
      "epoch": 24.6,
      "learning_rate": 3e-05,
      "loss": 0.49,
      "step": 8400
    },
    {
      "epoch": 24.89,
      "learning_rate": 3e-05,
      "loss": 0.5004,
      "step": 8500
    },
    {
      "epoch": 25.0,
      "eval_accuracy": 0.44035897435897436,
      "eval_loss": 4.332499027252197,
      "eval_runtime": 4.4052,
      "eval_samples_per_second": 113.503,
      "eval_steps_per_second": 14.301,
      "step": 8537
    },
    {
      "epoch": 25.0,
      "eval_bleu": 0.10401238605841524,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.024,
      "step": 8537
    },
    {
      "epoch": 25.18,
      "learning_rate": 3e-05,
      "loss": 0.4746,
      "step": 8600
    },
    {
      "epoch": 25.48,
      "learning_rate": 3e-05,
      "loss": 0.4596,
      "step": 8700
    },
    {
      "epoch": 25.77,
      "learning_rate": 3e-05,
      "loss": 0.4681,
      "step": 8800
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.43964102564102564,
      "eval_loss": 4.3162150382995605,
      "eval_runtime": 4.5151,
      "eval_samples_per_second": 110.739,
      "eval_steps_per_second": 13.953,
      "step": 8879
    },
    {
      "epoch": 26.0,
      "eval_bleu": 0.10710979645106765,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.026,
      "step": 8879
    },
    {
      "epoch": 26.06,
      "learning_rate": 3e-05,
      "loss": 0.459,
      "step": 8900
    },
    {
      "epoch": 26.35,
      "learning_rate": 3e-05,
      "loss": 0.4242,
      "step": 9000
    },
    {
      "epoch": 26.65,
      "learning_rate": 3e-05,
      "loss": 0.4445,
      "step": 9100
    },
    {
      "epoch": 26.94,
      "learning_rate": 3e-05,
      "loss": 0.4453,
      "step": 9200
    },
    {
      "epoch": 27.0,
      "eval_accuracy": 0.4388205128205128,
      "eval_loss": 4.377138137817383,
      "eval_runtime": 4.8206,
      "eval_samples_per_second": 103.722,
      "eval_steps_per_second": 13.069,
      "step": 9220
    },
    {
      "epoch": 27.0,
      "eval_bleu": 0.10010720174849254,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.026,
      "step": 9220
    },
    {
      "epoch": 27.23,
      "learning_rate": 3e-05,
      "loss": 0.4066,
      "step": 9300
    },
    {
      "epoch": 27.53,
      "learning_rate": 3e-05,
      "loss": 0.4153,
      "step": 9400
    },
    {
      "epoch": 27.82,
      "learning_rate": 3e-05,
      "loss": 0.4161,
      "step": 9500
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.43861538461538463,
      "eval_loss": 4.405981063842773,
      "eval_runtime": 4.4101,
      "eval_samples_per_second": 113.376,
      "eval_steps_per_second": 14.285,
      "step": 9562
    },
    {
      "epoch": 28.0,
      "eval_bleu": 0.10423428286785764,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.026,
      "step": 9562
    },
    {
      "epoch": 28.11,
      "learning_rate": 3e-05,
      "loss": 0.398,
      "step": 9600
    },
    {
      "epoch": 28.4,
      "learning_rate": 3e-05,
      "loss": 0.3788,
      "step": 9700
    },
    {
      "epoch": 28.7,
      "learning_rate": 3e-05,
      "loss": 0.3943,
      "step": 9800
    },
    {
      "epoch": 28.99,
      "learning_rate": 3e-05,
      "loss": 0.3994,
      "step": 9900
    },
    {
      "epoch": 29.0,
      "eval_accuracy": 0.4376923076923077,
      "eval_loss": 4.468777656555176,
      "eval_runtime": 4.2025,
      "eval_samples_per_second": 118.976,
      "eval_steps_per_second": 14.991,
      "step": 9903
    },
    {
      "epoch": 29.0,
      "eval_bleu": 0.08529933417800747,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.02,
      "step": 9903
    },
    {
      "epoch": 29.28,
      "learning_rate": 3e-05,
      "loss": 0.3519,
      "step": 10000
    },
    {
      "epoch": 29.58,
      "learning_rate": 3e-05,
      "loss": 0.3643,
      "step": 10100
    },
    {
      "epoch": 29.87,
      "learning_rate": 3e-05,
      "loss": 0.3695,
      "step": 10200
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.4376923076923077,
      "eval_loss": 4.464532375335693,
      "eval_runtime": 4.164,
      "eval_samples_per_second": 120.078,
      "eval_steps_per_second": 15.13,
      "step": 10245
    },
    {
      "epoch": 30.0,
      "eval_bleu": 0.08562517234751162,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.026,
      "step": 10245
    },
    {
      "epoch": 30.16,
      "learning_rate": 3e-05,
      "loss": 0.3474,
      "step": 10300
    },
    {
      "epoch": 30.45,
      "learning_rate": 3e-05,
      "loss": 0.3327,
      "step": 10400
    },
    {
      "epoch": 30.75,
      "learning_rate": 3e-05,
      "loss": 0.3505,
      "step": 10500
    },
    {
      "epoch": 31.0,
      "eval_accuracy": 0.4377948717948718,
      "eval_loss": 4.462403297424316,
      "eval_runtime": 4.2457,
      "eval_samples_per_second": 117.767,
      "eval_steps_per_second": 14.839,
      "step": 10586
    },
    {
      "epoch": 31.0,
      "eval_bleu": 0.09623161231398927,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.022,
      "step": 10586
    },
    {
      "epoch": 31.04,
      "learning_rate": 3e-05,
      "loss": 0.3476,
      "step": 10600
    },
    {
      "epoch": 31.33,
      "learning_rate": 3e-05,
      "loss": 0.3057,
      "step": 10700
    },
    {
      "epoch": 31.63,
      "learning_rate": 3e-05,
      "loss": 0.3224,
      "step": 10800
    },
    {
      "epoch": 31.92,
      "learning_rate": 3e-05,
      "loss": 0.3342,
      "step": 10900
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.43646153846153846,
      "eval_loss": 4.46300745010376,
      "eval_runtime": 4.5326,
      "eval_samples_per_second": 110.312,
      "eval_steps_per_second": 13.899,
      "step": 10928
    },
    {
      "epoch": 32.0,
      "eval_bleu": 0.09571435454261526,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.02,
      "step": 10928
    },
    {
      "epoch": 32.21,
      "learning_rate": 3e-05,
      "loss": 0.2943,
      "step": 11000
    },
    {
      "epoch": 32.5,
      "learning_rate": 3e-05,
      "loss": 0.3021,
      "step": 11100
    },
    {
      "epoch": 32.8,
      "learning_rate": 3e-05,
      "loss": 0.3075,
      "step": 11200
    },
    {
      "epoch": 33.0,
      "eval_accuracy": 0.4342051282051282,
      "eval_loss": 4.544379711151123,
      "eval_runtime": 4.4078,
      "eval_samples_per_second": 113.435,
      "eval_steps_per_second": 14.293,
      "step": 11269
    },
    {
      "epoch": 33.0,
      "eval_bleu": 0.11219607728661593,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.016,
      "step": 11269
    },
    {
      "epoch": 33.09,
      "learning_rate": 3e-05,
      "loss": 0.3047,
      "step": 11300
    },
    {
      "epoch": 33.38,
      "learning_rate": 3e-05,
      "loss": 0.2764,
      "step": 11400
    },
    {
      "epoch": 33.67,
      "learning_rate": 3e-05,
      "loss": 0.2833,
      "step": 11500
    },
    {
      "epoch": 33.97,
      "learning_rate": 3e-05,
      "loss": 0.2949,
      "step": 11600
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.43441025641025643,
      "eval_loss": 4.548118591308594,
      "eval_runtime": 4.4066,
      "eval_samples_per_second": 113.466,
      "eval_steps_per_second": 14.297,
      "step": 11611
    },
    {
      "epoch": 34.0,
      "eval_bleu": 0.10059053111753888,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.018,
      "step": 11611
    },
    {
      "epoch": 34.26,
      "learning_rate": 3e-05,
      "loss": 0.2604,
      "step": 11700
    },
    {
      "epoch": 34.55,
      "learning_rate": 3e-05,
      "loss": 0.2657,
      "step": 11800
    },
    {
      "epoch": 34.85,
      "learning_rate": 3e-05,
      "loss": 0.2705,
      "step": 11900
    },
    {
      "epoch": 35.0,
      "eval_accuracy": 0.4356923076923077,
      "eval_loss": 4.5613813400268555,
      "eval_runtime": 4.1998,
      "eval_samples_per_second": 119.055,
      "eval_steps_per_second": 15.001,
      "step": 11952
    },
    {
      "epoch": 35.0,
      "eval_bleu": 0.10287091260502411,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.012,
      "step": 11952
    },
    {
      "epoch": 35.14,
      "learning_rate": 3e-05,
      "loss": 0.2625,
      "step": 12000
    },
    {
      "epoch": 35.43,
      "learning_rate": 3e-05,
      "loss": 0.248,
      "step": 12100
    },
    {
      "epoch": 35.72,
      "learning_rate": 3e-05,
      "loss": 0.2554,
      "step": 12200
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.4339487179487179,
      "eval_loss": 4.591047763824463,
      "eval_runtime": 4.1092,
      "eval_samples_per_second": 121.677,
      "eval_steps_per_second": 15.331,
      "step": 12294
    },
    {
      "epoch": 36.0,
      "eval_bleu": 0.07688483668776659,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.02,
      "step": 12294
    },
    {
      "epoch": 36.02,
      "learning_rate": 3e-05,
      "loss": 0.2572,
      "step": 12300
    },
    {
      "epoch": 36.31,
      "learning_rate": 3e-05,
      "loss": 0.2276,
      "step": 12400
    },
    {
      "epoch": 36.6,
      "learning_rate": 3e-05,
      "loss": 0.236,
      "step": 12500
    },
    {
      "epoch": 36.9,
      "learning_rate": 3e-05,
      "loss": 0.2428,
      "step": 12600
    },
    {
      "epoch": 37.0,
      "eval_accuracy": 0.43323076923076925,
      "eval_loss": 4.645793914794922,
      "eval_runtime": 4.4058,
      "eval_samples_per_second": 113.486,
      "eval_steps_per_second": 14.299,
      "step": 12635
    },
    {
      "epoch": 37.0,
      "eval_bleu": 0.09755678571463763,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.01,
      "step": 12635
    },
    {
      "epoch": 37.19,
      "learning_rate": 3e-05,
      "loss": 0.2263,
      "step": 12700
    },
    {
      "epoch": 37.48,
      "learning_rate": 3e-05,
      "loss": 0.219,
      "step": 12800
    },
    {
      "epoch": 37.77,
      "learning_rate": 3e-05,
      "loss": 0.2277,
      "step": 12900
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.4327179487179487,
      "eval_loss": 4.655264377593994,
      "eval_runtime": 4.1048,
      "eval_samples_per_second": 121.809,
      "eval_steps_per_second": 15.348,
      "step": 12977
    },
    {
      "epoch": 38.0,
      "eval_bleu": 0.08910920223304278,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.016,
      "step": 12977
    },
    {
      "epoch": 38.07,
      "learning_rate": 3e-05,
      "loss": 0.2243,
      "step": 13000
    },
    {
      "epoch": 38.36,
      "learning_rate": 3e-05,
      "loss": 0.2037,
      "step": 13100
    },
    {
      "epoch": 38.65,
      "learning_rate": 3e-05,
      "loss": 0.2127,
      "step": 13200
    },
    {
      "epoch": 38.95,
      "learning_rate": 3e-05,
      "loss": 0.2172,
      "step": 13300
    },
    {
      "epoch": 39.0,
      "eval_accuracy": 0.4327692307692308,
      "eval_loss": 4.707071304321289,
      "eval_runtime": 4.4091,
      "eval_samples_per_second": 113.403,
      "eval_steps_per_second": 14.289,
      "step": 13318
    },
    {
      "epoch": 39.0,
      "eval_bleu": 0.08849488892310922,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.012,
      "step": 13318
    },
    {
      "epoch": 39.24,
      "learning_rate": 3e-05,
      "loss": 0.1976,
      "step": 13400
    },
    {
      "epoch": 39.53,
      "learning_rate": 3e-05,
      "loss": 0.1983,
      "step": 13500
    },
    {
      "epoch": 39.82,
      "learning_rate": 3e-05,
      "loss": 0.2016,
      "step": 13600
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.4330769230769231,
      "eval_loss": 4.717998504638672,
      "eval_runtime": 4.4131,
      "eval_samples_per_second": 113.298,
      "eval_steps_per_second": 14.276,
      "step": 13660
    },
    {
      "epoch": 40.0,
      "eval_bleu": 0.08922727862327968,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.018,
      "step": 13660
    },
    {
      "epoch": 40.12,
      "learning_rate": 3e-05,
      "loss": 0.1995,
      "step": 13700
    },
    {
      "epoch": 40.41,
      "learning_rate": 3e-05,
      "loss": 0.1852,
      "step": 13800
    },
    {
      "epoch": 40.7,
      "learning_rate": 3e-05,
      "loss": 0.1915,
      "step": 13900
    },
    {
      "epoch": 41.0,
      "learning_rate": 3e-05,
      "loss": 0.1965,
      "step": 14000
    },
    {
      "epoch": 41.0,
      "eval_accuracy": 0.43225641025641026,
      "eval_loss": 4.756761074066162,
      "eval_runtime": 4.223,
      "eval_samples_per_second": 118.4,
      "eval_steps_per_second": 14.918,
      "step": 14001
    },
    {
      "epoch": 41.0,
      "eval_bleu": 0.08298853800328854,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.01,
      "step": 14001
    },
    {
      "epoch": 41.29,
      "learning_rate": 3e-05,
      "loss": 0.1715,
      "step": 14100
    },
    {
      "epoch": 41.58,
      "learning_rate": 3e-05,
      "loss": 0.1816,
      "step": 14200
    },
    {
      "epoch": 41.87,
      "learning_rate": 3e-05,
      "loss": 0.1851,
      "step": 14300
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.4321025641025641,
      "eval_loss": 4.756207466125488,
      "eval_runtime": 4.826,
      "eval_samples_per_second": 103.604,
      "eval_steps_per_second": 13.054,
      "step": 14343
    },
    {
      "epoch": 42.0,
      "eval_bleu": 0.075425255663042,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.004,
      "step": 14343
    },
    {
      "epoch": 42.17,
      "learning_rate": 3e-05,
      "loss": 0.1704,
      "step": 14400
    },
    {
      "epoch": 42.46,
      "learning_rate": 3e-05,
      "loss": 0.1681,
      "step": 14500
    },
    {
      "epoch": 42.75,
      "learning_rate": 3e-05,
      "loss": 0.1739,
      "step": 14600
    },
    {
      "epoch": 43.0,
      "eval_accuracy": 0.4316923076923077,
      "eval_loss": 4.787448406219482,
      "eval_runtime": 4.1236,
      "eval_samples_per_second": 121.253,
      "eval_steps_per_second": 15.278,
      "step": 14684
    },
    {
      "epoch": 43.0,
      "eval_bleu": 0.08854177442021546,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.008,
      "step": 14684
    },
    {
      "epoch": 43.11,
      "learning_rate": 3e-05,
      "loss": 0.1439,
      "step": 14700
    },
    {
      "epoch": 43.4,
      "learning_rate": 3e-05,
      "loss": 0.1552,
      "step": 14800
    },
    {
      "epoch": 43.69,
      "learning_rate": 3e-05,
      "loss": 0.165,
      "step": 14900
    },
    {
      "epoch": 43.99,
      "learning_rate": 3e-05,
      "loss": 0.1719,
      "step": 15000
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.4323076923076923,
      "eval_loss": 4.802907943725586,
      "eval_runtime": 4.2242,
      "eval_samples_per_second": 118.365,
      "eval_steps_per_second": 14.914,
      "step": 15004
    },
    {
      "epoch": 44.0,
      "eval_bleu": 0.09013518216273568,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.006,
      "step": 15004
    },
    {
      "epoch": 44.28,
      "learning_rate": 3e-05,
      "loss": 0.1519,
      "step": 15100
    },
    {
      "epoch": 44.57,
      "learning_rate": 3e-05,
      "loss": 0.1562,
      "step": 15200
    },
    {
      "epoch": 44.87,
      "learning_rate": 3e-05,
      "loss": 0.1626,
      "step": 15300
    },
    {
      "epoch": 45.0,
      "eval_accuracy": 0.4318461538461538,
      "eval_loss": 4.781974792480469,
      "eval_runtime": 4.4138,
      "eval_samples_per_second": 113.282,
      "eval_steps_per_second": 14.274,
      "step": 15346
    },
    {
      "epoch": 45.0,
      "eval_bleu": 0.09066882849610237,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.01,
      "step": 15346
    },
    {
      "epoch": 45.16,
      "learning_rate": 3e-05,
      "loss": 0.1518,
      "step": 15400
    },
    {
      "epoch": 45.45,
      "learning_rate": 3e-05,
      "loss": 0.1465,
      "step": 15500
    },
    {
      "epoch": 45.74,
      "learning_rate": 3e-05,
      "loss": 0.1535,
      "step": 15600
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.43153846153846154,
      "eval_loss": 4.863694190979004,
      "eval_runtime": 4.8486,
      "eval_samples_per_second": 103.122,
      "eval_steps_per_second": 12.993,
      "step": 15687
    },
    {
      "epoch": 46.0,
      "eval_bleu": 0.08405923643052846,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.01,
      "step": 15687
    },
    {
      "epoch": 46.04,
      "learning_rate": 3e-05,
      "loss": 0.1567,
      "step": 15700
    },
    {
      "epoch": 46.33,
      "learning_rate": 3e-05,
      "loss": 0.1383,
      "step": 15800
    },
    {
      "epoch": 46.62,
      "learning_rate": 3e-05,
      "loss": 0.1454,
      "step": 15900
    },
    {
      "epoch": 46.92,
      "learning_rate": 3e-05,
      "loss": 0.1524,
      "step": 16000
    },
    {
      "epoch": 47.0,
      "eval_accuracy": 0.43148717948717946,
      "eval_loss": 4.89900016784668,
      "eval_runtime": 4.1156,
      "eval_samples_per_second": 121.488,
      "eval_steps_per_second": 15.307,
      "step": 16029
    },
    {
      "epoch": 47.0,
      "eval_bleu": 0.07603027545577727,
      "eval_exact_match": 0.0,
      "eval_prefix_exact_match": 0.01,
      "step": 16029
    }
  ],
  "logging_steps": 100,
  "max_steps": 17050,
  "num_train_epochs": 50,
  "save_steps": 500,
  "total_flos": 3.174158839649927e+18,
  "trial_name": null,
  "trial_params": null
}