{
  "best_metric": 0.6078237891197205,
  "best_model_checkpoint": "./whisper-tiny-tw/checkpoint-4000",
  "epoch": 5.333333333333333,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "learning_rate": 4.4e-07,
      "loss": 3.8753,
      "step": 25
    },
    {
      "epoch": 0.07,
      "learning_rate": 9.400000000000001e-07,
      "loss": 3.7563,
      "step": 50
    },
    {
      "epoch": 0.1,
      "learning_rate": 1.44e-06,
      "loss": 3.8738,
      "step": 75
    },
    {
      "epoch": 0.13,
      "learning_rate": 1.94e-06,
      "loss": 3.8201,
      "step": 100
    },
    {
      "epoch": 0.17,
      "learning_rate": 2.4400000000000004e-06,
      "loss": 3.8439,
      "step": 125
    },
    {
      "epoch": 0.2,
      "learning_rate": 2.9400000000000002e-06,
      "loss": 3.817,
      "step": 150
    },
    {
      "epoch": 0.23,
      "learning_rate": 3.44e-06,
      "loss": 3.8385,
      "step": 175
    },
    {
      "epoch": 0.27,
      "learning_rate": 3.94e-06,
      "loss": 3.9375,
      "step": 200
    },
    {
      "epoch": 0.3,
      "learning_rate": 4.440000000000001e-06,
      "loss": 3.8171,
      "step": 225
    },
    {
      "epoch": 0.33,
      "learning_rate": 4.94e-06,
      "loss": 3.8155,
      "step": 250
    },
    {
      "epoch": 0.37,
      "learning_rate": 5.4400000000000004e-06,
      "loss": 3.7763,
      "step": 275
    },
    {
      "epoch": 0.4,
      "learning_rate": 5.94e-06,
      "loss": 3.7395,
      "step": 300
    },
    {
      "epoch": 0.43,
      "learning_rate": 6.440000000000001e-06,
      "loss": 3.6626,
      "step": 325
    },
    {
      "epoch": 0.47,
      "learning_rate": 6.9400000000000005e-06,
      "loss": 3.6135,
      "step": 350
    },
    {
      "epoch": 0.5,
      "learning_rate": 7.440000000000001e-06,
      "loss": 3.6064,
      "step": 375
    },
    {
      "epoch": 0.53,
      "learning_rate": 7.94e-06,
      "loss": 3.6924,
      "step": 400
    },
    {
      "epoch": 0.57,
      "learning_rate": 8.44e-06,
      "loss": 3.5364,
      "step": 425
    },
    {
      "epoch": 0.6,
      "learning_rate": 8.94e-06,
      "loss": 3.621,
      "step": 450
    },
    {
      "epoch": 0.63,
      "learning_rate": 9.42e-06,
      "loss": 3.4745,
      "step": 475
    },
    {
      "epoch": 0.67,
      "learning_rate": 9.920000000000002e-06,
      "loss": 3.3802,
      "step": 500
    },
    {
      "epoch": 0.67,
      "eval_loss": 3.399221658706665,
      "eval_runtime": 217.7852,
      "eval_samples_per_second": 22.729,
      "eval_steps_per_second": 5.684,
      "step": 500
    },
    {
      "epoch": 0.7,
      "learning_rate": 9.940000000000001e-06,
      "loss": 3.3565,
      "step": 525
    },
    {
      "epoch": 0.73,
      "learning_rate": 9.86857142857143e-06,
      "loss": 3.2815,
      "step": 550
    },
    {
      "epoch": 0.77,
      "learning_rate": 9.797142857142858e-06,
      "loss": 3.252,
      "step": 575
    },
    {
      "epoch": 0.8,
      "learning_rate": 9.725714285714287e-06,
      "loss": 3.1434,
      "step": 600
    },
    {
      "epoch": 0.83,
      "learning_rate": 9.654285714285716e-06,
      "loss": 3.1253,
      "step": 625
    },
    {
      "epoch": 0.87,
      "learning_rate": 9.582857142857143e-06,
      "loss": 3.1106,
      "step": 650
    },
    {
      "epoch": 0.9,
      "learning_rate": 9.511428571428572e-06,
      "loss": 3.025,
      "step": 675
    },
    {
      "epoch": 0.93,
      "learning_rate": 9.440000000000001e-06,
      "loss": 2.9601,
      "step": 700
    },
    {
      "epoch": 0.97,
      "learning_rate": 9.368571428571428e-06,
      "loss": 2.8374,
      "step": 725
    },
    {
      "epoch": 1.0,
      "learning_rate": 9.297142857142857e-06,
      "loss": 2.7712,
      "step": 750
    },
    {
      "epoch": 1.03,
      "learning_rate": 9.225714285714286e-06,
      "loss": 2.6499,
      "step": 775
    },
    {
      "epoch": 1.07,
      "learning_rate": 9.154285714285715e-06,
      "loss": 2.6006,
      "step": 800
    },
    {
      "epoch": 1.1,
      "learning_rate": 9.082857142857143e-06,
      "loss": 2.5892,
      "step": 825
    },
    {
      "epoch": 1.13,
      "learning_rate": 9.011428571428572e-06,
      "loss": 2.4973,
      "step": 850
    },
    {
      "epoch": 1.17,
      "learning_rate": 8.94e-06,
      "loss": 2.4601,
      "step": 875
    },
    {
      "epoch": 1.2,
      "learning_rate": 8.86857142857143e-06,
      "loss": 2.3739,
      "step": 900
    },
    {
      "epoch": 1.23,
      "learning_rate": 8.797142857142857e-06,
      "loss": 2.3396,
      "step": 925
    },
    {
      "epoch": 1.27,
      "learning_rate": 8.725714285714286e-06,
      "loss": 2.3263,
      "step": 950
    },
    {
      "epoch": 1.3,
      "learning_rate": 8.654285714285715e-06,
      "loss": 2.2193,
      "step": 975
    },
    {
      "epoch": 1.33,
      "learning_rate": 8.582857142857144e-06,
      "loss": 2.1962,
      "step": 1000
    },
    {
      "epoch": 1.33,
      "eval_loss": 2.1643242835998535,
      "eval_runtime": 223.3057,
      "eval_samples_per_second": 22.167,
      "eval_steps_per_second": 5.544,
      "step": 1000
    },
    {
      "epoch": 1.37,
      "learning_rate": 8.511428571428571e-06,
      "loss": 2.1077,
      "step": 1025
    },
    {
      "epoch": 1.4,
      "learning_rate": 8.44e-06,
      "loss": 2.1055,
      "step": 1050
    },
    {
      "epoch": 1.43,
      "learning_rate": 8.36857142857143e-06,
      "loss": 2.0465,
      "step": 1075
    },
    {
      "epoch": 1.47,
      "learning_rate": 8.297142857142859e-06,
      "loss": 1.9805,
      "step": 1100
    },
    {
      "epoch": 1.5,
      "learning_rate": 8.225714285714288e-06,
      "loss": 2.0129,
      "step": 1125
    },
    {
      "epoch": 1.53,
      "learning_rate": 8.154285714285715e-06,
      "loss": 1.9447,
      "step": 1150
    },
    {
      "epoch": 1.57,
      "learning_rate": 8.082857142857144e-06,
      "loss": 1.9103,
      "step": 1175
    },
    {
      "epoch": 1.6,
      "learning_rate": 8.011428571428573e-06,
      "loss": 1.8697,
      "step": 1200
    },
    {
      "epoch": 1.63,
      "learning_rate": 7.94e-06,
      "loss": 1.8626,
      "step": 1225
    },
    {
      "epoch": 1.67,
      "learning_rate": 7.86857142857143e-06,
      "loss": 1.8127,
      "step": 1250
    },
    {
      "epoch": 1.7,
      "learning_rate": 7.797142857142858e-06,
      "loss": 1.8026,
      "step": 1275
    },
    {
      "epoch": 1.73,
      "learning_rate": 7.725714285714286e-06,
      "loss": 1.7526,
      "step": 1300
    },
    {
      "epoch": 1.77,
      "learning_rate": 7.654285714285715e-06,
      "loss": 1.6291,
      "step": 1325
    },
    {
      "epoch": 1.8,
      "learning_rate": 7.5828571428571444e-06,
      "loss": 1.6233,
      "step": 1350
    },
    {
      "epoch": 1.83,
      "learning_rate": 7.511428571428572e-06,
      "loss": 1.6377,
      "step": 1375
    },
    {
      "epoch": 1.87,
      "learning_rate": 7.440000000000001e-06,
      "loss": 1.5758,
      "step": 1400
    },
    {
      "epoch": 1.9,
      "learning_rate": 7.36857142857143e-06,
      "loss": 1.5483,
      "step": 1425
    },
    {
      "epoch": 1.93,
      "learning_rate": 7.297142857142858e-06,
      "loss": 1.5166,
      "step": 1450
    },
    {
      "epoch": 1.97,
      "learning_rate": 7.225714285714286e-06,
      "loss": 1.4319,
      "step": 1475
    },
    {
      "epoch": 2.0,
      "learning_rate": 7.154285714285715e-06,
      "loss": 1.4348,
      "step": 1500
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.4068281650543213,
      "eval_runtime": 219.1085,
      "eval_samples_per_second": 22.592,
      "eval_steps_per_second": 5.65,
      "step": 1500
    },
    {
      "epoch": 2.03,
      "learning_rate": 7.082857142857143e-06,
      "loss": 1.3035,
      "step": 1525
    },
    {
      "epoch": 2.07,
      "learning_rate": 7.011428571428572e-06,
      "loss": 1.3554,
      "step": 1550
    },
    {
      "epoch": 2.1,
      "learning_rate": 6.9400000000000005e-06,
      "loss": 1.3143,
      "step": 1575
    },
    {
      "epoch": 2.13,
      "learning_rate": 6.868571428571429e-06,
      "loss": 1.2939,
      "step": 1600
    },
    {
      "epoch": 2.17,
      "learning_rate": 6.797142857142858e-06,
      "loss": 1.2099,
      "step": 1625
    },
    {
      "epoch": 2.2,
      "learning_rate": 6.725714285714287e-06,
      "loss": 1.2248,
      "step": 1650
    },
    {
      "epoch": 2.23,
      "learning_rate": 6.654285714285716e-06,
      "loss": 1.1217,
      "step": 1675
    },
    {
      "epoch": 2.27,
      "learning_rate": 6.582857142857143e-06,
      "loss": 1.1274,
      "step": 1700
    },
    {
      "epoch": 2.3,
      "learning_rate": 6.511428571428572e-06,
      "loss": 1.0399,
      "step": 1725
    },
    {
      "epoch": 2.33,
      "learning_rate": 6.440000000000001e-06,
      "loss": 1.046,
      "step": 1750
    },
    {
      "epoch": 2.37,
      "learning_rate": 6.371428571428572e-06,
      "loss": 0.9206,
      "step": 1775
    },
    {
      "epoch": 2.4,
      "learning_rate": 6.300000000000001e-06,
      "loss": 0.9125,
      "step": 1800
    },
    {
      "epoch": 2.43,
      "learning_rate": 6.22857142857143e-06,
      "loss": 0.9636,
      "step": 1825
    },
    {
      "epoch": 2.47,
      "learning_rate": 6.157142857142858e-06,
      "loss": 0.8434,
      "step": 1850
    },
    {
      "epoch": 2.5,
      "learning_rate": 6.085714285714286e-06,
      "loss": 0.7642,
      "step": 1875
    },
    {
      "epoch": 2.53,
      "learning_rate": 6.014285714285715e-06,
      "loss": 0.7749,
      "step": 1900
    },
    {
      "epoch": 2.57,
      "learning_rate": 5.942857142857143e-06,
      "loss": 0.7122,
      "step": 1925
    },
    {
      "epoch": 2.6,
      "learning_rate": 5.871428571428572e-06,
      "loss": 0.7065,
      "step": 1950
    },
    {
      "epoch": 2.63,
      "learning_rate": 5.8e-06,
      "loss": 0.7062,
      "step": 1975
    },
    {
      "epoch": 2.67,
      "learning_rate": 5.7285714285714285e-06,
      "loss": 0.7108,
      "step": 2000
    },
    {
      "epoch": 2.67,
      "eval_loss": 0.6925591230392456,
      "eval_runtime": 219.8232,
      "eval_samples_per_second": 22.518,
      "eval_steps_per_second": 5.632,
      "step": 2000
    },
    {
      "epoch": 2.7,
      "learning_rate": 5.6571428571428576e-06,
      "loss": 0.7476,
      "step": 2025
    },
    {
      "epoch": 2.73,
      "learning_rate": 5.5857142857142866e-06,
      "loss": 0.6809,
      "step": 2050
    },
    {
      "epoch": 2.77,
      "learning_rate": 5.514285714285714e-06,
      "loss": 0.625,
      "step": 2075
    },
    {
      "epoch": 2.8,
      "learning_rate": 5.442857142857143e-06,
      "loss": 0.6623,
      "step": 2100
    },
    {
      "epoch": 2.83,
      "learning_rate": 5.371428571428572e-06,
      "loss": 0.7041,
      "step": 2125
    },
    {
      "epoch": 2.87,
      "learning_rate": 5.300000000000001e-06,
      "loss": 0.6609,
      "step": 2150
    },
    {
      "epoch": 2.9,
      "learning_rate": 5.22857142857143e-06,
      "loss": 0.7054,
      "step": 2175
    },
    {
      "epoch": 2.93,
      "learning_rate": 5.157142857142857e-06,
      "loss": 0.694,
      "step": 2200
    },
    {
      "epoch": 2.97,
      "learning_rate": 5.085714285714286e-06,
      "loss": 0.702,
      "step": 2225
    },
    {
      "epoch": 3.0,
      "learning_rate": 5.014285714285715e-06,
      "loss": 0.6537,
      "step": 2250
    },
    {
      "epoch": 3.03,
      "learning_rate": 4.9428571428571435e-06,
      "loss": 0.6367,
      "step": 2275
    },
    {
      "epoch": 3.07,
      "learning_rate": 4.871428571428572e-06,
      "loss": 0.6481,
      "step": 2300
    },
    {
      "epoch": 3.1,
      "learning_rate": 4.800000000000001e-06,
      "loss": 0.6755,
      "step": 2325
    },
    {
      "epoch": 3.13,
      "learning_rate": 4.728571428571429e-06,
      "loss": 0.6257,
      "step": 2350
    },
    {
      "epoch": 3.17,
      "learning_rate": 4.657142857142857e-06,
      "loss": 0.6438,
      "step": 2375
    },
    {
      "epoch": 3.2,
      "learning_rate": 4.585714285714286e-06,
      "loss": 0.7029,
      "step": 2400
    },
    {
      "epoch": 3.23,
      "learning_rate": 4.514285714285714e-06,
      "loss": 0.6453,
      "step": 2425
    },
    {
      "epoch": 3.27,
      "learning_rate": 4.442857142857143e-06,
      "loss": 0.6728,
      "step": 2450
    },
    {
      "epoch": 3.3,
      "learning_rate": 4.371428571428572e-06,
      "loss": 0.6667,
      "step": 2475
    },
    {
      "epoch": 3.33,
      "learning_rate": 4.3e-06,
      "loss": 0.6801,
      "step": 2500
    },
    {
      "epoch": 3.33,
      "eval_loss": 0.6374496221542358,
      "eval_runtime": 222.1867,
      "eval_samples_per_second": 22.279,
      "eval_steps_per_second": 5.572,
      "step": 2500
    },
    {
      "epoch": 3.37,
      "learning_rate": 4.228571428571429e-06,
      "loss": 0.6021,
      "step": 2525
    },
    {
      "epoch": 3.4,
      "learning_rate": 4.1571428571428575e-06,
      "loss": 0.6657,
      "step": 2550
    },
    {
      "epoch": 3.43,
      "learning_rate": 4.0857142857142865e-06,
      "loss": 0.6338,
      "step": 2575
    },
    {
      "epoch": 3.47,
      "learning_rate": 4.014285714285715e-06,
      "loss": 0.6559,
      "step": 2600
    },
    {
      "epoch": 3.5,
      "learning_rate": 3.942857142857143e-06,
      "loss": 0.6559,
      "step": 2625
    },
    {
      "epoch": 3.53,
      "learning_rate": 3.871428571428572e-06,
      "loss": 0.6894,
      "step": 2650
    },
    {
      "epoch": 3.57,
      "learning_rate": 3.8000000000000005e-06,
      "loss": 0.6422,
      "step": 2675
    },
    {
      "epoch": 3.6,
      "learning_rate": 3.7285714285714286e-06,
      "loss": 0.6211,
      "step": 2700
    },
    {
      "epoch": 3.63,
      "learning_rate": 3.6571428571428576e-06,
      "loss": 0.6228,
      "step": 2725
    },
    {
      "epoch": 3.67,
      "learning_rate": 3.5857142857142862e-06,
      "loss": 0.5956,
      "step": 2750
    },
    {
      "epoch": 3.7,
      "learning_rate": 3.5142857142857144e-06,
      "loss": 0.6002,
      "step": 2775
    },
    {
      "epoch": 3.73,
      "learning_rate": 3.4428571428571434e-06,
      "loss": 0.6384,
      "step": 2800
    },
    {
      "epoch": 3.77,
      "learning_rate": 3.3714285714285716e-06,
      "loss": 0.6283,
      "step": 2825
    },
    {
      "epoch": 3.8,
      "learning_rate": 3.3000000000000006e-06,
      "loss": 0.6691,
      "step": 2850
    },
    {
      "epoch": 3.83,
      "learning_rate": 3.2285714285714288e-06,
      "loss": 0.5512,
      "step": 2875
    },
    {
      "epoch": 3.87,
      "learning_rate": 3.1571428571428573e-06,
      "loss": 0.5573,
      "step": 2900
    },
    {
      "epoch": 3.9,
      "learning_rate": 3.085714285714286e-06,
      "loss": 0.5772,
      "step": 2925
    },
    {
      "epoch": 3.93,
      "learning_rate": 3.0142857142857145e-06,
      "loss": 0.6307,
      "step": 2950
    },
    {
      "epoch": 3.97,
      "learning_rate": 2.9428571428571427e-06,
      "loss": 0.5967,
      "step": 2975
    },
    {
      "epoch": 4.0,
      "learning_rate": 2.8714285714285717e-06,
      "loss": 0.6273,
      "step": 3000
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.6194672584533691,
      "eval_runtime": 222.9187,
      "eval_samples_per_second": 22.205,
      "eval_steps_per_second": 5.554,
      "step": 3000
    },
    {
      "epoch": 4.03,
      "learning_rate": 2.8000000000000003e-06,
      "loss": 0.5915,
      "step": 3025
    },
    {
      "epoch": 4.07,
      "learning_rate": 2.728571428571429e-06,
      "loss": 0.6411,
      "step": 3050
    },
    {
      "epoch": 4.1,
      "learning_rate": 2.6571428571428575e-06,
      "loss": 0.6522,
      "step": 3075
    },
    {
      "epoch": 4.13,
      "learning_rate": 2.5857142857142856e-06,
      "loss": 0.5862,
      "step": 3100
    },
    {
      "epoch": 4.17,
      "learning_rate": 2.5142857142857147e-06,
      "loss": 0.5907,
      "step": 3125
    },
    {
      "epoch": 4.2,
      "learning_rate": 2.442857142857143e-06,
      "loss": 0.6228,
      "step": 3150
    },
    {
      "epoch": 4.23,
      "learning_rate": 2.371428571428572e-06,
      "loss": 0.639,
      "step": 3175
    },
    {
      "epoch": 4.27,
      "learning_rate": 2.3000000000000004e-06,
      "loss": 0.6105,
      "step": 3200
    },
    {
      "epoch": 4.3,
      "learning_rate": 2.228571428571429e-06,
      "loss": 0.6154,
      "step": 3225
    },
    {
      "epoch": 4.33,
      "learning_rate": 2.157142857142857e-06,
      "loss": 0.6179,
      "step": 3250
    },
    {
      "epoch": 4.37,
      "learning_rate": 2.0857142857142858e-06,
      "loss": 0.5826,
      "step": 3275
    },
    {
      "epoch": 4.4,
      "learning_rate": 2.0142857142857144e-06,
      "loss": 0.6063,
      "step": 3300
    },
    {
      "epoch": 4.43,
      "learning_rate": 1.942857142857143e-06,
      "loss": 0.6412,
      "step": 3325
    },
    {
      "epoch": 4.47,
      "learning_rate": 1.8714285714285715e-06,
      "loss": 0.5746,
      "step": 3350
    },
    {
      "epoch": 4.5,
      "learning_rate": 1.8000000000000001e-06,
      "loss": 0.6644,
      "step": 3375
    },
    {
      "epoch": 4.53,
      "learning_rate": 1.7285714285714287e-06,
      "loss": 0.5814,
      "step": 3400
    },
    {
      "epoch": 4.57,
      "learning_rate": 1.657142857142857e-06,
      "loss": 0.6841,
      "step": 3425
    },
    {
      "epoch": 4.6,
      "learning_rate": 1.5857142857142857e-06,
      "loss": 0.6066,
      "step": 3450
    },
    {
      "epoch": 4.63,
      "learning_rate": 1.5142857142857145e-06,
      "loss": 0.5992,
      "step": 3475
    },
    {
      "epoch": 4.67,
      "learning_rate": 1.442857142857143e-06,
      "loss": 0.6001,
      "step": 3500
    },
    {
      "epoch": 4.67,
      "eval_loss": 0.6105530858039856,
      "eval_runtime": 216.5509,
      "eval_samples_per_second": 22.858,
      "eval_steps_per_second": 5.717,
      "step": 3500
    },
    {
      "epoch": 4.7,
      "learning_rate": 1.3714285714285717e-06,
      "loss": 0.6524,
      "step": 3525
    },
    {
      "epoch": 4.73,
      "learning_rate": 1.3e-06,
      "loss": 0.5826,
      "step": 3550
    },
    {
      "epoch": 4.77,
      "learning_rate": 1.2285714285714286e-06,
      "loss": 0.5674,
      "step": 3575
    },
    {
      "epoch": 4.8,
      "learning_rate": 1.1571428571428572e-06,
      "loss": 0.5817,
      "step": 3600
    },
    {
      "epoch": 4.83,
      "learning_rate": 1.0857142857142858e-06,
      "loss": 0.6095,
      "step": 3625
    },
    {
      "epoch": 4.87,
      "learning_rate": 1.0142857142857144e-06,
      "loss": 0.6374,
      "step": 3650
    },
    {
      "epoch": 4.9,
      "learning_rate": 9.42857142857143e-07,
      "loss": 0.6038,
      "step": 3675
    },
    {
      "epoch": 4.93,
      "learning_rate": 8.714285714285716e-07,
      "loss": 0.6349,
      "step": 3700
    },
    {
      "epoch": 4.97,
      "learning_rate": 8.000000000000001e-07,
      "loss": 0.6017,
      "step": 3725
    },
    {
      "epoch": 5.0,
      "learning_rate": 7.285714285714287e-07,
      "loss": 0.6295,
      "step": 3750
    },
    {
      "epoch": 5.03,
      "learning_rate": 6.571428571428571e-07,
      "loss": 0.631,
      "step": 3775
    },
    {
      "epoch": 5.07,
      "learning_rate": 5.857142857142857e-07,
      "loss": 0.5753,
      "step": 3800
    },
    {
      "epoch": 5.1,
      "learning_rate": 5.142857142857143e-07,
      "loss": 0.5888,
      "step": 3825
    },
    {
      "epoch": 5.13,
      "learning_rate": 4.4285714285714286e-07,
      "loss": 0.6269,
      "step": 3850
    },
    {
      "epoch": 5.17,
      "learning_rate": 3.7142857142857145e-07,
      "loss": 0.5584,
      "step": 3875
    },
    {
      "epoch": 5.2,
      "learning_rate": 3.0000000000000004e-07,
      "loss": 0.642,
      "step": 3900
    },
    {
      "epoch": 5.23,
      "learning_rate": 2.285714285714286e-07,
      "loss": 0.561,
      "step": 3925
    },
    {
      "epoch": 5.27,
      "learning_rate": 1.5714285714285717e-07,
      "loss": 0.6469,
      "step": 3950
    },
    {
      "epoch": 5.3,
      "learning_rate": 8.571428571428573e-08,
      "loss": 0.6164,
      "step": 3975
    },
    {
      "epoch": 5.33,
      "learning_rate": 1.4285714285714288e-08,
      "loss": 0.6082,
      "step": 4000
    },
    {
      "epoch": 5.33,
      "eval_loss": 0.6078237891197205,
      "eval_runtime": 220.2328,
      "eval_samples_per_second": 22.476,
      "eval_steps_per_second": 5.621,
      "step": 4000
    },
    {
      "epoch": 5.33,
      "step": 4000,
      "total_flos": 1.6501699427328e+18,
      "train_loss": 1.4726774854660034,
      "train_runtime": 5301.1107,
      "train_samples_per_second": 12.073,
      "train_steps_per_second": 0.755
    }
  ],
  "logging_steps": 25,
  "max_steps": 4000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 6,
  "save_steps": 500,
  "total_flos": 1.6501699427328e+18,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}