{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 98.2905982905983,
  "global_step": 23000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 2.14,
      "learning_rate": 4.8933760683760685e-05,
      "loss": 1.1035,
      "step": 500
    },
    {
      "epoch": 2.14,
      "eval_loss": 0.7537529468536377,
      "eval_runtime": 164.1555,
      "eval_samples_per_second": 1072.605,
      "eval_steps_per_second": 33.523,
      "step": 500
    },
    {
      "epoch": 4.27,
      "learning_rate": 4.7865384615384615e-05,
      "loss": 0.7267,
      "step": 1000
    },
    {
      "epoch": 4.27,
      "eval_loss": 0.6768979430198669,
      "eval_runtime": 180.9639,
      "eval_samples_per_second": 972.979,
      "eval_steps_per_second": 30.409,
      "step": 1000
    },
    {
      "epoch": 6.41,
      "learning_rate": 4.6797008547008546e-05,
      "loss": 0.6744,
      "step": 1500
    },
    {
      "epoch": 6.41,
      "eval_loss": 0.6488664746284485,
      "eval_runtime": 190.376,
      "eval_samples_per_second": 924.875,
      "eval_steps_per_second": 28.906,
      "step": 1500
    },
    {
      "epoch": 8.55,
      "learning_rate": 4.572863247863248e-05,
      "loss": 0.6491,
      "step": 2000
    },
    {
      "epoch": 8.55,
      "eval_loss": 0.6347554326057434,
      "eval_runtime": 166.0432,
      "eval_samples_per_second": 1060.411,
      "eval_steps_per_second": 33.142,
      "step": 2000
    },
    {
      "epoch": 10.68,
      "learning_rate": 4.466025641025641e-05,
      "loss": 0.6334,
      "step": 2500
    },
    {
      "epoch": 10.68,
      "eval_loss": 0.6249234080314636,
      "eval_runtime": 165.6169,
      "eval_samples_per_second": 1063.141,
      "eval_steps_per_second": 33.227,
      "step": 2500
    },
    {
      "epoch": 12.82,
      "learning_rate": 4.359188034188034e-05,
      "loss": 0.6219,
      "step": 3000
    },
    {
      "epoch": 12.82,
      "eval_loss": 0.6181825995445251,
      "eval_runtime": 167.5877,
      "eval_samples_per_second": 1050.638,
      "eval_steps_per_second": 32.837,
      "step": 3000
    },
    {
      "epoch": 14.96,
      "learning_rate": 4.252350427350428e-05,
      "loss": 0.6131,
      "step": 3500
    },
    {
      "epoch": 14.96,
      "eval_loss": 0.6141465902328491,
      "eval_runtime": 177.8868,
      "eval_samples_per_second": 989.809,
      "eval_steps_per_second": 30.935,
      "step": 3500
    },
    {
      "epoch": 17.09,
      "learning_rate": 4.145512820512821e-05,
      "loss": 0.6058,
      "step": 4000
    },
    {
      "epoch": 17.09,
      "eval_loss": 0.6131681203842163,
      "eval_runtime": 163.4313,
      "eval_samples_per_second": 1077.358,
      "eval_steps_per_second": 33.672,
      "step": 4000
    },
    {
      "epoch": 19.23,
      "learning_rate": 4.0386752136752134e-05,
      "loss": 0.5995,
      "step": 4500
    },
    {
      "epoch": 19.23,
      "eval_loss": 0.6094414591789246,
      "eval_runtime": 178.3809,
      "eval_samples_per_second": 987.068,
      "eval_steps_per_second": 30.85,
      "step": 4500
    },
    {
      "epoch": 21.37,
      "learning_rate": 3.931837606837607e-05,
      "loss": 0.5941,
      "step": 5000
    },
    {
      "epoch": 21.37,
      "eval_loss": 0.6074140667915344,
      "eval_runtime": 169.9857,
      "eval_samples_per_second": 1035.817,
      "eval_steps_per_second": 32.373,
      "step": 5000
    },
    {
      "epoch": 23.5,
      "learning_rate": 3.825e-05,
      "loss": 0.5894,
      "step": 5500
    },
    {
      "epoch": 23.5,
      "eval_loss": 0.6062588691711426,
      "eval_runtime": 168.4324,
      "eval_samples_per_second": 1045.369,
      "eval_steps_per_second": 32.672,
      "step": 5500
    },
    {
      "epoch": 25.64,
      "learning_rate": 3.718162393162393e-05,
      "loss": 0.5851,
      "step": 6000
    },
    {
      "epoch": 25.64,
      "eval_loss": 0.6049907207489014,
      "eval_runtime": 169.165,
      "eval_samples_per_second": 1040.842,
      "eval_steps_per_second": 32.53,
      "step": 6000
    },
    {
      "epoch": 27.78,
      "learning_rate": 3.611324786324787e-05,
      "loss": 0.5813,
      "step": 6500
    },
    {
      "epoch": 27.78,
      "eval_loss": 0.6046043634414673,
      "eval_runtime": 164.2895,
      "eval_samples_per_second": 1071.73,
      "eval_steps_per_second": 33.496,
      "step": 6500
    },
    {
      "epoch": 29.91,
      "learning_rate": 3.50448717948718e-05,
      "loss": 0.5777,
      "step": 7000
    },
    {
      "epoch": 29.91,
      "eval_loss": 0.6043672561645508,
      "eval_runtime": 163.9442,
      "eval_samples_per_second": 1073.987,
      "eval_steps_per_second": 33.566,
      "step": 7000
    },
    {
      "epoch": 32.05,
      "learning_rate": 3.397649572649573e-05,
      "loss": 0.5744,
      "step": 7500
    },
    {
      "epoch": 32.05,
      "eval_loss": 0.6073368787765503,
      "eval_runtime": 182.0186,
      "eval_samples_per_second": 967.341,
      "eval_steps_per_second": 30.233,
      "step": 7500
    },
    {
      "epoch": 34.19,
      "learning_rate": 3.290811965811966e-05,
      "loss": 0.5712,
      "step": 8000
    },
    {
      "epoch": 34.19,
      "eval_loss": 0.6060159802436829,
      "eval_runtime": 164.583,
      "eval_samples_per_second": 1069.819,
      "eval_steps_per_second": 33.436,
      "step": 8000
    },
    {
      "epoch": 36.32,
      "learning_rate": 3.183974358974359e-05,
      "loss": 0.5684,
      "step": 8500
    },
    {
      "epoch": 36.32,
      "eval_loss": 0.6061359643936157,
      "eval_runtime": 165.5065,
      "eval_samples_per_second": 1063.849,
      "eval_steps_per_second": 33.249,
      "step": 8500
    },
    {
      "epoch": 38.46,
      "learning_rate": 3.077136752136752e-05,
      "loss": 0.5658,
      "step": 9000
    },
    {
      "epoch": 38.46,
      "eval_loss": 0.6064005494117737,
      "eval_runtime": 177.3855,
      "eval_samples_per_second": 992.607,
      "eval_steps_per_second": 31.023,
      "step": 9000
    },
    {
      "epoch": 40.6,
      "learning_rate": 2.9702991452991454e-05,
      "loss": 0.5634,
      "step": 9500
    },
    {
      "epoch": 40.6,
      "eval_loss": 0.6069360375404358,
      "eval_runtime": 164.0377,
      "eval_samples_per_second": 1073.375,
      "eval_steps_per_second": 33.547,
      "step": 9500
    },
    {
      "epoch": 42.74,
      "learning_rate": 2.8634615384615387e-05,
      "loss": 0.5612,
      "step": 10000
    },
    {
      "epoch": 42.74,
      "eval_loss": 0.6074803471565247,
      "eval_runtime": 184.5732,
      "eval_samples_per_second": 953.952,
      "eval_steps_per_second": 29.815,
      "step": 10000
    },
    {
      "epoch": 44.87,
      "learning_rate": 2.7566239316239318e-05,
      "loss": 0.5591,
      "step": 10500
    },
    {
      "epoch": 44.87,
      "eval_loss": 0.6080661416053772,
      "eval_runtime": 164.771,
      "eval_samples_per_second": 1068.598,
      "eval_steps_per_second": 33.398,
      "step": 10500
    },
    {
      "epoch": 47.01,
      "learning_rate": 2.6497863247863248e-05,
      "loss": 0.5571,
      "step": 11000
    },
    {
      "epoch": 47.01,
      "eval_loss": 0.610197901725769,
      "eval_runtime": 163.8142,
      "eval_samples_per_second": 1074.84,
      "eval_steps_per_second": 33.593,
      "step": 11000
    },
    {
      "epoch": 49.15,
      "learning_rate": 2.542948717948718e-05,
      "loss": 0.5549,
      "step": 11500
    },
    {
      "epoch": 49.15,
      "eval_loss": 0.6111814379692078,
      "eval_runtime": 172.4133,
      "eval_samples_per_second": 1021.232,
      "eval_steps_per_second": 31.917,
      "step": 11500
    },
    {
      "epoch": 51.28,
      "learning_rate": 2.4361111111111112e-05,
      "loss": 0.5532,
      "step": 12000
    },
    {
      "epoch": 51.28,
      "eval_loss": 0.611818790435791,
      "eval_runtime": 174.1127,
      "eval_samples_per_second": 1011.264,
      "eval_steps_per_second": 31.606,
      "step": 12000
    },
    {
      "epoch": 53.42,
      "learning_rate": 2.3292735042735045e-05,
      "loss": 0.5515,
      "step": 12500
    },
    {
      "epoch": 53.42,
      "eval_loss": 0.6132317781448364,
      "eval_runtime": 164.6297,
      "eval_samples_per_second": 1069.515,
      "eval_steps_per_second": 33.427,
      "step": 12500
    },
    {
      "epoch": 55.56,
      "learning_rate": 2.2224358974358976e-05,
      "loss": 0.55,
      "step": 13000
    },
    {
      "epoch": 55.56,
      "eval_loss": 0.6136063933372498,
      "eval_runtime": 164.5555,
      "eval_samples_per_second": 1069.998,
      "eval_steps_per_second": 33.442,
      "step": 13000
    },
    {
      "epoch": 57.69,
      "learning_rate": 2.1155982905982906e-05,
      "loss": 0.5486,
      "step": 13500
    },
    {
      "epoch": 57.69,
      "eval_loss": 0.6144915819168091,
      "eval_runtime": 203.4153,
      "eval_samples_per_second": 865.589,
      "eval_steps_per_second": 27.053,
      "step": 13500
    },
    {
      "epoch": 59.83,
      "learning_rate": 2.008760683760684e-05,
      "loss": 0.5472,
      "step": 14000
    },
    {
      "epoch": 59.83,
      "eval_loss": 0.6147069931030273,
      "eval_runtime": 199.2649,
      "eval_samples_per_second": 883.618,
      "eval_steps_per_second": 27.617,
      "step": 14000
    },
    {
      "epoch": 61.97,
      "learning_rate": 1.9021367521367522e-05,
      "loss": 0.5457,
      "step": 14500
    },
    {
      "epoch": 61.97,
      "eval_loss": 0.6158111095428467,
      "eval_runtime": 206.6245,
      "eval_samples_per_second": 852.145,
      "eval_steps_per_second": 26.633,
      "step": 14500
    },
    {
      "epoch": 64.1,
      "learning_rate": 1.7952991452991456e-05,
      "loss": 0.5445,
      "step": 15000
    },
    {
      "epoch": 64.1,
      "eval_loss": 0.6180585026741028,
      "eval_runtime": 181.3159,
      "eval_samples_per_second": 971.09,
      "eval_steps_per_second": 30.35,
      "step": 15000
    },
    {
      "epoch": 66.24,
      "learning_rate": 1.688675213675214e-05,
      "loss": 0.5431,
      "step": 15500
    },
    {
      "epoch": 66.24,
      "eval_loss": 0.6186466217041016,
      "eval_runtime": 175.3115,
      "eval_samples_per_second": 1004.349,
      "eval_steps_per_second": 31.39,
      "step": 15500
    },
    {
      "epoch": 68.38,
      "learning_rate": 1.581837606837607e-05,
      "loss": 0.5421,
      "step": 16000
    },
    {
      "epoch": 68.38,
      "eval_loss": 0.6195328831672668,
      "eval_runtime": 164.9282,
      "eval_samples_per_second": 1067.58,
      "eval_steps_per_second": 33.366,
      "step": 16000
    },
    {
      "epoch": 70.51,
      "learning_rate": 1.475e-05,
      "loss": 0.5411,
      "step": 16500
    },
    {
      "epoch": 70.51,
      "eval_loss": 0.6201679110527039,
      "eval_runtime": 197.124,
      "eval_samples_per_second": 893.215,
      "eval_steps_per_second": 27.916,
      "step": 16500
    },
    {
      "epoch": 72.65,
      "learning_rate": 1.3681623931623933e-05,
      "loss": 0.5401,
      "step": 17000
    },
    {
      "epoch": 72.65,
      "eval_loss": 0.6210618019104004,
      "eval_runtime": 173.1765,
      "eval_samples_per_second": 1016.732,
      "eval_steps_per_second": 31.777,
      "step": 17000
    },
    {
      "epoch": 74.79,
      "learning_rate": 1.2613247863247865e-05,
      "loss": 0.5391,
      "step": 17500
    },
    {
      "epoch": 74.79,
      "eval_loss": 0.6214122772216797,
      "eval_runtime": 196.4031,
      "eval_samples_per_second": 896.493,
      "eval_steps_per_second": 28.019,
      "step": 17500
    },
    {
      "epoch": 76.92,
      "learning_rate": 1.1544871794871795e-05,
      "loss": 0.5382,
      "step": 18000
    },
    {
      "epoch": 76.92,
      "eval_loss": 0.6220324635505676,
      "eval_runtime": 164.5331,
      "eval_samples_per_second": 1070.143,
      "eval_steps_per_second": 33.446,
      "step": 18000
    },
    {
      "epoch": 79.06,
      "learning_rate": 1.0476495726495727e-05,
      "loss": 0.5372,
      "step": 18500
    },
    {
      "epoch": 79.06,
      "eval_loss": 0.6238572001457214,
      "eval_runtime": 180.654,
      "eval_samples_per_second": 974.647,
      "eval_steps_per_second": 30.462,
      "step": 18500
    },
    {
      "epoch": 81.2,
      "learning_rate": 9.40811965811966e-06,
      "loss": 0.5363,
      "step": 19000
    },
    {
      "epoch": 81.2,
      "eval_loss": 0.6245262026786804,
      "eval_runtime": 163.6759,
      "eval_samples_per_second": 1075.748,
      "eval_steps_per_second": 33.621,
      "step": 19000
    },
    {
      "epoch": 83.33,
      "learning_rate": 8.341880341880342e-06,
      "loss": 0.5355,
      "step": 19500
    },
    {
      "epoch": 83.33,
      "eval_loss": 0.6251608729362488,
      "eval_runtime": 200.2426,
      "eval_samples_per_second": 879.303,
      "eval_steps_per_second": 27.482,
      "step": 19500
    },
    {
      "epoch": 85.47,
      "learning_rate": 7.273504273504274e-06,
      "loss": 0.5348,
      "step": 20000
    },
    {
      "epoch": 85.47,
      "eval_loss": 0.6253054738044739,
      "eval_runtime": 175.0233,
      "eval_samples_per_second": 1006.003,
      "eval_steps_per_second": 31.442,
      "step": 20000
    },
    {
      "epoch": 87.61,
      "learning_rate": 6.205128205128206e-06,
      "loss": 0.5341,
      "step": 20500
    },
    {
      "epoch": 87.61,
      "eval_loss": 0.6260212063789368,
      "eval_runtime": 165.1088,
      "eval_samples_per_second": 1066.412,
      "eval_steps_per_second": 33.33,
      "step": 20500
    },
    {
      "epoch": 89.74,
      "learning_rate": 5.136752136752137e-06,
      "loss": 0.5334,
      "step": 21000
    },
    {
      "epoch": 89.74,
      "eval_loss": 0.6266852021217346,
      "eval_runtime": 164.357,
      "eval_samples_per_second": 1071.29,
      "eval_steps_per_second": 33.482,
      "step": 21000
    },
    {
      "epoch": 91.88,
      "learning_rate": 4.068376068376069e-06,
      "loss": 0.5327,
      "step": 21500
    },
    {
      "epoch": 91.88,
      "eval_loss": 0.6268437504768372,
      "eval_runtime": 168.7818,
      "eval_samples_per_second": 1043.205,
      "eval_steps_per_second": 32.604,
      "step": 21500
    },
    {
      "epoch": 94.02,
      "learning_rate": 3e-06,
      "loss": 0.5321,
      "step": 22000
    },
    {
      "epoch": 94.02,
      "eval_loss": 0.6274420022964478,
      "eval_runtime": 193.8022,
      "eval_samples_per_second": 908.524,
      "eval_steps_per_second": 28.395,
      "step": 22000
    },
    {
      "epoch": 96.15,
      "learning_rate": 1.9316239316239316e-06,
      "loss": 0.5315,
      "step": 22500
    },
    {
      "epoch": 96.15,
      "eval_loss": 0.6278537511825562,
      "eval_runtime": 195.5578,
      "eval_samples_per_second": 900.368,
      "eval_steps_per_second": 28.14,
      "step": 22500
    },
    {
      "epoch": 98.29,
      "learning_rate": 8.632478632478633e-07,
      "loss": 0.531,
      "step": 23000
    },
    {
      "epoch": 98.29,
      "eval_loss": 0.6279672980308533,
      "eval_runtime": 167.092,
      "eval_samples_per_second": 1053.755,
      "eval_steps_per_second": 32.934,
      "step": 23000
    }
  ],
  "max_steps": 23400,
  "num_train_epochs": 100,
  "total_flos": 2.3846898985745777e+18,
  "trial_name": null,
  "trial_params": null
}