{
  "best_metric": 95.5,
  "best_model_checkpoint": "./medium_TH/checkpoint-3000",
  "epoch": 88.23529411764706,
  "global_step": 3000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.74,
      "learning_rate": 4.4e-07,
      "loss": 0.8659,
      "step": 25
    },
    {
      "epoch": 1.47,
      "learning_rate": 9.400000000000001e-07,
      "loss": 0.7368,
      "step": 50
    },
    {
      "epoch": 2.21,
      "learning_rate": 1.44e-06,
      "loss": 0.5972,
      "step": 75
    },
    {
      "epoch": 2.94,
      "learning_rate": 1.94e-06,
      "loss": 0.5492,
      "step": 100
    },
    {
      "epoch": 3.68,
      "learning_rate": 2.4400000000000004e-06,
      "loss": 0.4563,
      "step": 125
    },
    {
      "epoch": 4.41,
      "learning_rate": 2.9400000000000002e-06,
      "loss": 0.4365,
      "step": 150
    },
    {
      "epoch": 5.15,
      "learning_rate": 3.44e-06,
      "loss": 0.5196,
      "step": 175
    },
    {
      "epoch": 5.88,
      "learning_rate": 3.94e-06,
      "loss": 0.3682,
      "step": 200
    },
    {
      "epoch": 6.62,
      "learning_rate": 4.440000000000001e-06,
      "loss": 0.2693,
      "step": 225
    },
    {
      "epoch": 7.35,
      "learning_rate": 4.94e-06,
      "loss": 0.2433,
      "step": 250
    },
    {
      "epoch": 8.09,
      "learning_rate": 5.4400000000000004e-06,
      "loss": 0.2578,
      "step": 275
    },
    {
      "epoch": 8.82,
      "learning_rate": 5.94e-06,
      "loss": 0.173,
      "step": 300
    },
    {
      "epoch": 9.56,
      "learning_rate": 6.440000000000001e-06,
      "loss": 0.1173,
      "step": 325
    },
    {
      "epoch": 10.29,
      "learning_rate": 6.9400000000000005e-06,
      "loss": 0.1491,
      "step": 350
    },
    {
      "epoch": 11.03,
      "learning_rate": 7.440000000000001e-06,
      "loss": 0.1102,
      "step": 375
    },
    {
      "epoch": 11.76,
      "learning_rate": 7.94e-06,
      "loss": 0.1325,
      "step": 400
    },
    {
      "epoch": 12.5,
      "learning_rate": 8.44e-06,
      "loss": 0.1195,
      "step": 425
    },
    {
      "epoch": 13.24,
      "learning_rate": 8.94e-06,
      "loss": 0.0834,
      "step": 450
    },
    {
      "epoch": 13.97,
      "learning_rate": 9.440000000000001e-06,
      "loss": 0.1089,
      "step": 475
    },
    {
      "epoch": 14.71,
      "learning_rate": 9.940000000000001e-06,
      "loss": 0.0594,
      "step": 500
    },
    {
      "epoch": 15.44,
      "learning_rate": 9.937142857142858e-06,
      "loss": 0.0973,
      "step": 525
    },
    {
      "epoch": 16.18,
      "learning_rate": 9.865714285714285e-06,
      "loss": 0.0344,
      "step": 550
    },
    {
      "epoch": 16.91,
      "learning_rate": 9.794285714285714e-06,
      "loss": 0.0628,
      "step": 575
    },
    {
      "epoch": 17.65,
      "learning_rate": 9.722857142857143e-06,
      "loss": 0.0611,
      "step": 600
    },
    {
      "epoch": 18.38,
      "learning_rate": 9.651428571428572e-06,
      "loss": 0.054,
      "step": 625
    },
    {
      "epoch": 19.12,
      "learning_rate": 9.58e-06,
      "loss": 0.069,
      "step": 650
    },
    {
      "epoch": 19.85,
      "learning_rate": 9.508571428571429e-06,
      "loss": 0.0491,
      "step": 675
    },
    {
      "epoch": 20.59,
      "learning_rate": 9.437142857142858e-06,
      "loss": 0.0958,
      "step": 700
    },
    {
      "epoch": 21.32,
      "learning_rate": 9.365714285714287e-06,
      "loss": 0.0477,
      "step": 725
    },
    {
      "epoch": 22.06,
      "learning_rate": 9.294285714285714e-06,
      "loss": 0.0318,
      "step": 750
    },
    {
      "epoch": 22.79,
      "learning_rate": 9.222857142857143e-06,
      "loss": 0.0691,
      "step": 775
    },
    {
      "epoch": 23.53,
      "learning_rate": 9.151428571428572e-06,
      "loss": 0.0317,
      "step": 800
    },
    {
      "epoch": 24.26,
      "learning_rate": 9.080000000000001e-06,
      "loss": 0.0221,
      "step": 825
    },
    {
      "epoch": 25.0,
      "learning_rate": 9.00857142857143e-06,
      "loss": 0.039,
      "step": 850
    },
    {
      "epoch": 25.74,
      "learning_rate": 8.937142857142857e-06,
      "loss": 0.0344,
      "step": 875
    },
    {
      "epoch": 26.47,
      "learning_rate": 8.865714285714287e-06,
      "loss": 0.0168,
      "step": 900
    },
    {
      "epoch": 27.21,
      "learning_rate": 8.794285714285716e-06,
      "loss": 0.0346,
      "step": 925
    },
    {
      "epoch": 27.94,
      "learning_rate": 8.722857142857145e-06,
      "loss": 0.0437,
      "step": 950
    },
    {
      "epoch": 28.68,
      "learning_rate": 8.651428571428572e-06,
      "loss": 0.0197,
      "step": 975
    },
    {
      "epoch": 29.41,
      "learning_rate": 8.580000000000001e-06,
      "loss": 0.0303,
      "step": 1000
    },
    {
      "epoch": 29.41,
      "eval_loss": 0.3286643922328949,
      "eval_runtime": 10.3915,
      "eval_samples_per_second": 2.213,
      "eval_steps_per_second": 0.289,
      "eval_wer": 106.0,
      "step": 1000
    },
    {
      "epoch": 30.15,
      "learning_rate": 8.50857142857143e-06,
      "loss": 0.0517,
      "step": 1025
    },
    {
      "epoch": 30.88,
      "learning_rate": 8.437142857142859e-06,
      "loss": 0.0381,
      "step": 1050
    },
    {
      "epoch": 31.62,
      "learning_rate": 8.365714285714286e-06,
      "loss": 0.0309,
      "step": 1075
    },
    {
      "epoch": 32.35,
      "learning_rate": 8.294285714285715e-06,
      "loss": 0.0078,
      "step": 1100
    },
    {
      "epoch": 33.09,
      "learning_rate": 8.222857142857144e-06,
      "loss": 0.0297,
      "step": 1125
    },
    {
      "epoch": 33.82,
      "learning_rate": 8.151428571428572e-06,
      "loss": 0.0247,
      "step": 1150
    },
    {
      "epoch": 34.56,
      "learning_rate": 8.08e-06,
      "loss": 0.0098,
      "step": 1175
    },
    {
      "epoch": 35.29,
      "learning_rate": 8.00857142857143e-06,
      "loss": 0.0083,
      "step": 1200
    },
    {
      "epoch": 36.03,
      "learning_rate": 7.937142857142857e-06,
      "loss": 0.0296,
      "step": 1225
    },
    {
      "epoch": 36.76,
      "learning_rate": 7.865714285714286e-06,
      "loss": 0.021,
      "step": 1250
    },
    {
      "epoch": 37.5,
      "learning_rate": 7.794285714285715e-06,
      "loss": 0.0403,
      "step": 1275
    },
    {
      "epoch": 38.24,
      "learning_rate": 7.722857142857142e-06,
      "loss": 0.0225,
      "step": 1300
    },
    {
      "epoch": 38.97,
      "learning_rate": 7.651428571428571e-06,
      "loss": 0.0295,
      "step": 1325
    },
    {
      "epoch": 39.71,
      "learning_rate": 7.58e-06,
      "loss": 0.0144,
      "step": 1350
    },
    {
      "epoch": 40.44,
      "learning_rate": 7.508571428571429e-06,
      "loss": 0.0271,
      "step": 1375
    },
    {
      "epoch": 41.18,
      "learning_rate": 7.4371428571428575e-06,
      "loss": 0.0225,
      "step": 1400
    },
    {
      "epoch": 41.91,
      "learning_rate": 7.365714285714286e-06,
      "loss": 0.0242,
      "step": 1425
    },
    {
      "epoch": 42.65,
      "learning_rate": 7.294285714285715e-06,
      "loss": 0.0138,
      "step": 1450
    },
    {
      "epoch": 43.38,
      "learning_rate": 7.222857142857144e-06,
      "loss": 0.0106,
      "step": 1475
    },
    {
      "epoch": 44.12,
      "learning_rate": 7.151428571428573e-06,
      "loss": 0.0221,
      "step": 1500
    },
    {
      "epoch": 44.85,
      "learning_rate": 7.08e-06,
      "loss": 0.0304,
      "step": 1525
    },
    {
      "epoch": 45.59,
      "learning_rate": 7.008571428571429e-06,
      "loss": 0.0224,
      "step": 1550
    },
    {
      "epoch": 46.32,
      "learning_rate": 6.937142857142858e-06,
      "loss": 0.0208,
      "step": 1575
    },
    {
      "epoch": 47.06,
      "learning_rate": 6.865714285714287e-06,
      "loss": 0.0057,
      "step": 1600
    },
    {
      "epoch": 47.79,
      "learning_rate": 6.794285714285714e-06,
      "loss": 0.0181,
      "step": 1625
    },
    {
      "epoch": 48.53,
      "learning_rate": 6.722857142857143e-06,
      "loss": 0.0118,
      "step": 1650
    },
    {
      "epoch": 49.26,
      "learning_rate": 6.651428571428572e-06,
      "loss": 0.0158,
      "step": 1675
    },
    {
      "epoch": 50.0,
      "learning_rate": 6.5800000000000005e-06,
      "loss": 0.015,
      "step": 1700
    },
    {
      "epoch": 50.74,
      "learning_rate": 6.5085714285714295e-06,
      "loss": 0.0095,
      "step": 1725
    },
    {
      "epoch": 51.47,
      "learning_rate": 6.437142857142858e-06,
      "loss": 0.015,
      "step": 1750
    },
    {
      "epoch": 52.21,
      "learning_rate": 6.365714285714286e-06,
      "loss": 0.0099,
      "step": 1775
    },
    {
      "epoch": 52.94,
      "learning_rate": 6.294285714285715e-06,
      "loss": 0.0111,
      "step": 1800
    },
    {
      "epoch": 53.68,
      "learning_rate": 6.222857142857144e-06,
      "loss": 0.0165,
      "step": 1825
    },
    {
      "epoch": 54.41,
      "learning_rate": 6.151428571428571e-06,
      "loss": 0.0097,
      "step": 1850
    },
    {
      "epoch": 55.15,
      "learning_rate": 6.08e-06,
      "loss": 0.0042,
      "step": 1875
    },
    {
      "epoch": 55.88,
      "learning_rate": 6.008571428571429e-06,
      "loss": 0.0222,
      "step": 1900
    },
    {
      "epoch": 56.62,
      "learning_rate": 5.937142857142858e-06,
      "loss": 0.009,
      "step": 1925
    },
    {
      "epoch": 57.35,
      "learning_rate": 5.865714285714286e-06,
      "loss": 0.0052,
      "step": 1950
    },
    {
      "epoch": 58.09,
      "learning_rate": 5.794285714285715e-06,
      "loss": 0.0126,
      "step": 1975
    },
    {
      "epoch": 58.82,
      "learning_rate": 5.722857142857144e-06,
      "loss": 0.0142,
      "step": 2000
    },
    {
      "epoch": 58.82,
      "eval_loss": 0.3790155053138733,
      "eval_runtime": 10.2927,
      "eval_samples_per_second": 2.235,
      "eval_steps_per_second": 0.291,
      "eval_wer": 99.5,
      "step": 2000
    },
    {
      "epoch": 59.56,
      "learning_rate": 5.651428571428572e-06,
      "loss": 0.0117,
      "step": 2025
    },
    {
      "epoch": 60.29,
      "learning_rate": 5.580000000000001e-06,
      "loss": 0.0079,
      "step": 2050
    },
    {
      "epoch": 61.03,
      "learning_rate": 5.508571428571429e-06,
      "loss": 0.0044,
      "step": 2075
    },
    {
      "epoch": 61.76,
      "learning_rate": 5.437142857142857e-06,
      "loss": 0.007,
      "step": 2100
    },
    {
      "epoch": 62.5,
      "learning_rate": 5.365714285714286e-06,
      "loss": 0.021,
      "step": 2125
    },
    {
      "epoch": 63.24,
      "learning_rate": 5.294285714285715e-06,
      "loss": 0.0102,
      "step": 2150
    },
    {
      "epoch": 63.97,
      "learning_rate": 5.2228571428571425e-06,
      "loss": 0.0035,
      "step": 2175
    },
    {
      "epoch": 64.71,
      "learning_rate": 5.1514285714285715e-06,
      "loss": 0.0072,
      "step": 2200
    },
    {
      "epoch": 65.44,
      "learning_rate": 5.0800000000000005e-06,
      "loss": 0.0032,
      "step": 2225
    },
    {
      "epoch": 66.18,
      "learning_rate": 5.0085714285714295e-06,
      "loss": 0.0047,
      "step": 2250
    },
    {
      "epoch": 66.91,
      "learning_rate": 4.937142857142858e-06,
      "loss": 0.003,
      "step": 2275
    },
    {
      "epoch": 67.65,
      "learning_rate": 4.865714285714287e-06,
      "loss": 0.0114,
      "step": 2300
    },
    {
      "epoch": 68.38,
      "learning_rate": 4.794285714285715e-06,
      "loss": 0.0073,
      "step": 2325
    },
    {
      "epoch": 69.12,
      "learning_rate": 4.722857142857144e-06,
      "loss": 0.0119,
      "step": 2350
    },
    {
      "epoch": 69.85,
      "learning_rate": 4.651428571428572e-06,
      "loss": 0.0045,
      "step": 2375
    },
    {
      "epoch": 70.59,
      "learning_rate": 4.58e-06,
      "loss": 0.0147,
      "step": 2400
    },
    {
      "epoch": 71.32,
      "learning_rate": 4.508571428571429e-06,
      "loss": 0.0081,
      "step": 2425
    },
    {
      "epoch": 72.06,
      "learning_rate": 4.437142857142857e-06,
      "loss": 0.0159,
      "step": 2450
    },
    {
      "epoch": 72.79,
      "learning_rate": 4.3657142857142855e-06,
      "loss": 0.0032,
      "step": 2475
    },
    {
      "epoch": 73.53,
      "learning_rate": 4.2942857142857146e-06,
      "loss": 0.0081,
      "step": 2500
    },
    {
      "epoch": 74.26,
      "learning_rate": 4.222857142857143e-06,
      "loss": 0.0116,
      "step": 2525
    },
    {
      "epoch": 75.0,
      "learning_rate": 4.151428571428572e-06,
      "loss": 0.0012,
      "step": 2550
    },
    {
      "epoch": 75.74,
      "learning_rate": 4.08e-06,
      "loss": 0.0018,
      "step": 2575
    },
    {
      "epoch": 76.47,
      "learning_rate": 4.008571428571429e-06,
      "loss": 0.0059,
      "step": 2600
    },
    {
      "epoch": 77.21,
      "learning_rate": 3.937142857142858e-06,
      "loss": 0.0032,
      "step": 2625
    },
    {
      "epoch": 77.94,
      "learning_rate": 3.865714285714286e-06,
      "loss": 0.0089,
      "step": 2650
    },
    {
      "epoch": 78.68,
      "learning_rate": 3.7942857142857147e-06,
      "loss": 0.0044,
      "step": 2675
    },
    {
      "epoch": 79.41,
      "learning_rate": 3.722857142857143e-06,
      "loss": 0.0049,
      "step": 2700
    },
    {
      "epoch": 80.15,
      "learning_rate": 3.651428571428572e-06,
      "loss": 0.0038,
      "step": 2725
    },
    {
      "epoch": 80.88,
      "learning_rate": 3.58e-06,
      "loss": 0.0053,
      "step": 2750
    },
    {
      "epoch": 81.62,
      "learning_rate": 3.508571428571429e-06,
      "loss": 0.004,
      "step": 2775
    },
    {
      "epoch": 82.35,
      "learning_rate": 3.437142857142857e-06,
      "loss": 0.0495,
      "step": 2800
    },
    {
      "epoch": 83.09,
      "learning_rate": 3.3657142857142862e-06,
      "loss": 0.0035,
      "step": 2825
    },
    {
      "epoch": 83.82,
      "learning_rate": 3.2942857142857144e-06,
      "loss": 0.0044,
      "step": 2850
    },
    {
      "epoch": 84.56,
      "learning_rate": 3.222857142857143e-06,
      "loss": 0.005,
      "step": 2875
    },
    {
      "epoch": 85.29,
      "learning_rate": 3.151428571428572e-06,
      "loss": 0.0102,
      "step": 2900
    },
    {
      "epoch": 86.03,
      "learning_rate": 3.08e-06,
      "loss": 0.0054,
      "step": 2925
    },
    {
      "epoch": 86.76,
      "learning_rate": 3.008571428571429e-06,
      "loss": 0.0095,
      "step": 2950
    },
    {
      "epoch": 87.5,
      "learning_rate": 2.9371428571428573e-06,
      "loss": 0.0065,
      "step": 2975
    },
    {
      "epoch": 88.24,
      "learning_rate": 2.865714285714286e-06,
      "loss": 0.0032,
      "step": 3000
    },
    {
      "epoch": 88.24,
      "eval_loss": 0.3970164954662323,
      "eval_runtime": 10.2993,
      "eval_samples_per_second": 2.233,
      "eval_steps_per_second": 0.291,
      "eval_wer": 95.5,
      "step": 3000
    }
  ],
  "max_steps": 4000,
  "num_train_epochs": 118,
  "total_flos": 1.801163540791296e+19,
  "trial_name": null,
  "trial_params": null
}