{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.5195263290501386,
  "eval_steps": 500,
  "global_step": 20000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.06,
      "learning_rate": 0.000975,
      "loss": 1.0238,
      "step": 500
    },
    {
      "epoch": 0.06,
      "eval_loss": 0.884382426738739,
      "eval_runtime": 4.1887,
      "eval_samples_per_second": 238.736,
      "eval_steps_per_second": 29.842,
      "step": 500
    },
    {
      "epoch": 0.13,
      "learning_rate": 0.00095,
      "loss": 0.9971,
      "step": 1000
    },
    {
      "epoch": 0.13,
      "eval_loss": 0.8562089800834656,
      "eval_runtime": 4.1399,
      "eval_samples_per_second": 241.551,
      "eval_steps_per_second": 30.194,
      "step": 1000
    },
    {
      "epoch": 0.19,
      "learning_rate": 0.000925,
      "loss": 0.9522,
      "step": 1500
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.8645691871643066,
      "eval_runtime": 4.1915,
      "eval_samples_per_second": 238.576,
      "eval_steps_per_second": 29.822,
      "step": 1500
    },
    {
      "epoch": 0.25,
      "learning_rate": 0.0009000000000000001,
      "loss": 0.9143,
      "step": 2000
    },
    {
      "epoch": 0.25,
      "eval_loss": 0.8668216466903687,
      "eval_runtime": 4.1753,
      "eval_samples_per_second": 239.502,
      "eval_steps_per_second": 29.938,
      "step": 2000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000875,
      "loss": 0.9151,
      "step": 2500
    },
    {
      "epoch": 0.31,
      "eval_loss": 0.8562004566192627,
      "eval_runtime": 4.2561,
      "eval_samples_per_second": 234.956,
      "eval_steps_per_second": 29.369,
      "step": 2500
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.00085,
      "loss": 0.9263,
      "step": 3000
    },
    {
      "epoch": 0.38,
      "eval_loss": 0.850217342376709,
      "eval_runtime": 4.1404,
      "eval_samples_per_second": 241.521,
      "eval_steps_per_second": 30.19,
      "step": 3000
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.000825,
      "loss": 0.9102,
      "step": 3500
    },
    {
      "epoch": 0.44,
      "eval_loss": 0.8581109642982483,
      "eval_runtime": 4.1878,
      "eval_samples_per_second": 238.788,
      "eval_steps_per_second": 29.849,
      "step": 3500
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.0008,
      "loss": 0.8866,
      "step": 4000
    },
    {
      "epoch": 0.5,
      "eval_loss": 0.8308799862861633,
      "eval_runtime": 4.1108,
      "eval_samples_per_second": 243.259,
      "eval_steps_per_second": 30.407,
      "step": 4000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.0007750000000000001,
      "loss": 0.911,
      "step": 4500
    },
    {
      "epoch": 0.57,
      "eval_loss": 0.801396369934082,
      "eval_runtime": 4.1725,
      "eval_samples_per_second": 239.665,
      "eval_steps_per_second": 29.958,
      "step": 4500
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.00075,
      "loss": 0.8536,
      "step": 5000
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.8181760907173157,
      "eval_runtime": 4.1738,
      "eval_samples_per_second": 239.59,
      "eval_steps_per_second": 29.949,
      "step": 5000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.000725,
      "loss": 0.8277,
      "step": 5500
    },
    {
      "epoch": 0.69,
      "eval_loss": 0.7971529960632324,
      "eval_runtime": 4.1414,
      "eval_samples_per_second": 241.466,
      "eval_steps_per_second": 30.183,
      "step": 5500
    },
    {
      "epoch": 0.76,
      "learning_rate": 0.0007,
      "loss": 0.8413,
      "step": 6000
    },
    {
      "epoch": 0.76,
      "eval_loss": 0.8046051263809204,
      "eval_runtime": 4.278,
      "eval_samples_per_second": 233.755,
      "eval_steps_per_second": 29.219,
      "step": 6000
    },
    {
      "epoch": 0.82,
      "learning_rate": 0.000675,
      "loss": 0.8491,
      "step": 6500
    },
    {
      "epoch": 0.82,
      "eval_loss": 0.8008071184158325,
      "eval_runtime": 4.2965,
      "eval_samples_per_second": 232.746,
      "eval_steps_per_second": 29.093,
      "step": 6500
    },
    {
      "epoch": 0.88,
      "learning_rate": 0.0006500000000000001,
      "loss": 0.8077,
      "step": 7000
    },
    {
      "epoch": 0.88,
      "eval_loss": 0.7875821590423584,
      "eval_runtime": 4.1627,
      "eval_samples_per_second": 240.231,
      "eval_steps_per_second": 30.029,
      "step": 7000
    },
    {
      "epoch": 0.94,
      "learning_rate": 0.000625,
      "loss": 0.796,
      "step": 7500
    },
    {
      "epoch": 0.94,
      "eval_loss": 0.7883646488189697,
      "eval_runtime": 4.2129,
      "eval_samples_per_second": 237.368,
      "eval_steps_per_second": 29.671,
      "step": 7500
    },
    {
      "epoch": 1.01,
      "learning_rate": 0.0006,
      "loss": 0.7952,
      "step": 8000
    },
    {
      "epoch": 1.01,
      "eval_loss": 0.7713276743888855,
      "eval_runtime": 4.1008,
      "eval_samples_per_second": 243.858,
      "eval_steps_per_second": 30.482,
      "step": 8000
    },
    {
      "epoch": 1.07,
      "learning_rate": 0.000575,
      "loss": 0.5761,
      "step": 8500
    },
    {
      "epoch": 1.07,
      "eval_loss": 0.7843192219734192,
      "eval_runtime": 4.1527,
      "eval_samples_per_second": 240.808,
      "eval_steps_per_second": 30.101,
      "step": 8500
    },
    {
      "epoch": 1.13,
      "learning_rate": 0.00055,
      "loss": 0.6084,
      "step": 9000
    },
    {
      "epoch": 1.13,
      "eval_loss": 0.7597091197967529,
      "eval_runtime": 4.2348,
      "eval_samples_per_second": 236.137,
      "eval_steps_per_second": 29.517,
      "step": 9000
    },
    {
      "epoch": 1.2,
      "learning_rate": 0.0005250000000000001,
      "loss": 0.5863,
      "step": 9500
    },
    {
      "epoch": 1.2,
      "eval_loss": 0.7753661274909973,
      "eval_runtime": 4.2434,
      "eval_samples_per_second": 235.658,
      "eval_steps_per_second": 29.457,
      "step": 9500
    },
    {
      "epoch": 1.26,
      "learning_rate": 0.0005,
      "loss": 0.5835,
      "step": 10000
    },
    {
      "epoch": 1.26,
      "eval_loss": 0.7525186538696289,
      "eval_runtime": 4.1329,
      "eval_samples_per_second": 241.962,
      "eval_steps_per_second": 30.245,
      "step": 10000
    },
    {
      "epoch": 1.32,
      "learning_rate": 0.000475,
      "loss": 0.5751,
      "step": 10500
    },
    {
      "epoch": 1.32,
      "eval_loss": 0.7624223828315735,
      "eval_runtime": 4.1841,
      "eval_samples_per_second": 239.002,
      "eval_steps_per_second": 29.875,
      "step": 10500
    },
    {
      "epoch": 1.39,
      "learning_rate": 0.00045000000000000004,
      "loss": 0.5746,
      "step": 11000
    },
    {
      "epoch": 1.39,
      "eval_loss": 0.7706524729728699,
      "eval_runtime": 4.2408,
      "eval_samples_per_second": 235.806,
      "eval_steps_per_second": 29.476,
      "step": 11000
    },
    {
      "epoch": 1.45,
      "learning_rate": 0.000425,
      "loss": 0.5847,
      "step": 11500
    },
    {
      "epoch": 1.45,
      "eval_loss": 0.7473410367965698,
      "eval_runtime": 4.1712,
      "eval_samples_per_second": 239.742,
      "eval_steps_per_second": 29.968,
      "step": 11500
    },
    {
      "epoch": 1.51,
      "learning_rate": 0.0004,
      "loss": 0.5759,
      "step": 12000
    },
    {
      "epoch": 1.51,
      "eval_loss": 0.7421715259552002,
      "eval_runtime": 4.1644,
      "eval_samples_per_second": 240.128,
      "eval_steps_per_second": 30.016,
      "step": 12000
    },
    {
      "epoch": 1.57,
      "learning_rate": 0.000375,
      "loss": 0.5922,
      "step": 12500
    },
    {
      "epoch": 1.57,
      "eval_loss": 0.7362275123596191,
      "eval_runtime": 4.1889,
      "eval_samples_per_second": 238.726,
      "eval_steps_per_second": 29.841,
      "step": 12500
    },
    {
      "epoch": 1.64,
      "learning_rate": 0.00035,
      "loss": 0.5678,
      "step": 13000
    },
    {
      "epoch": 1.64,
      "eval_loss": 0.7327093482017517,
      "eval_runtime": 4.151,
      "eval_samples_per_second": 240.907,
      "eval_steps_per_second": 30.113,
      "step": 13000
    },
    {
      "epoch": 1.7,
      "learning_rate": 0.00032500000000000004,
      "loss": 0.5627,
      "step": 13500
    },
    {
      "epoch": 1.7,
      "eval_loss": 0.7284647226333618,
      "eval_runtime": 4.2289,
      "eval_samples_per_second": 236.47,
      "eval_steps_per_second": 29.559,
      "step": 13500
    },
    {
      "epoch": 1.76,
      "learning_rate": 0.0003,
      "loss": 0.5568,
      "step": 14000
    },
    {
      "epoch": 1.76,
      "eval_loss": 0.7164832949638367,
      "eval_runtime": 4.2221,
      "eval_samples_per_second": 236.848,
      "eval_steps_per_second": 29.606,
      "step": 14000
    },
    {
      "epoch": 1.83,
      "learning_rate": 0.000275,
      "loss": 0.5589,
      "step": 14500
    },
    {
      "epoch": 1.83,
      "eval_loss": 0.7153368592262268,
      "eval_runtime": 4.2183,
      "eval_samples_per_second": 237.06,
      "eval_steps_per_second": 29.633,
      "step": 14500
    },
    {
      "epoch": 1.89,
      "learning_rate": 0.00025,
      "loss": 0.5675,
      "step": 15000
    },
    {
      "epoch": 1.89,
      "eval_loss": 0.7135257720947266,
      "eval_runtime": 4.2335,
      "eval_samples_per_second": 236.21,
      "eval_steps_per_second": 29.526,
      "step": 15000
    },
    {
      "epoch": 1.95,
      "learning_rate": 0.00022500000000000002,
      "loss": 0.5422,
      "step": 15500
    },
    {
      "epoch": 1.95,
      "eval_loss": 0.719129741191864,
      "eval_runtime": 4.2149,
      "eval_samples_per_second": 237.255,
      "eval_steps_per_second": 29.657,
      "step": 15500
    },
    {
      "epoch": 2.02,
      "learning_rate": 0.0002,
      "loss": 0.5104,
      "step": 16000
    },
    {
      "epoch": 2.02,
      "eval_loss": 0.7474156022071838,
      "eval_runtime": 4.1473,
      "eval_samples_per_second": 241.123,
      "eval_steps_per_second": 30.14,
      "step": 16000
    },
    {
      "epoch": 2.08,
      "learning_rate": 0.000175,
      "loss": 0.3835,
      "step": 16500
    },
    {
      "epoch": 2.08,
      "eval_loss": 0.7469050288200378,
      "eval_runtime": 4.1498,
      "eval_samples_per_second": 240.978,
      "eval_steps_per_second": 30.122,
      "step": 16500
    },
    {
      "epoch": 2.14,
      "learning_rate": 0.00015,
      "loss": 0.3805,
      "step": 17000
    },
    {
      "epoch": 2.14,
      "eval_loss": 0.7489848136901855,
      "eval_runtime": 3.9927,
      "eval_samples_per_second": 250.456,
      "eval_steps_per_second": 31.307,
      "step": 17000
    },
    {
      "epoch": 2.2,
      "learning_rate": 0.000125,
      "loss": 0.3718,
      "step": 17500
    },
    {
      "epoch": 2.2,
      "eval_loss": 0.7569396495819092,
      "eval_runtime": 4.0483,
      "eval_samples_per_second": 247.019,
      "eval_steps_per_second": 30.877,
      "step": 17500
    },
    {
      "epoch": 2.27,
      "learning_rate": 0.0001,
      "loss": 0.3688,
      "step": 18000
    },
    {
      "epoch": 2.27,
      "eval_loss": 0.7492154836654663,
      "eval_runtime": 4.0676,
      "eval_samples_per_second": 245.848,
      "eval_steps_per_second": 30.731,
      "step": 18000
    },
    {
      "epoch": 2.33,
      "learning_rate": 7.5e-05,
      "loss": 0.3645,
      "step": 18500
    },
    {
      "epoch": 2.33,
      "eval_loss": 0.7465029954910278,
      "eval_runtime": 4.2652,
      "eval_samples_per_second": 234.456,
      "eval_steps_per_second": 29.307,
      "step": 18500
    },
    {
      "epoch": 2.39,
      "learning_rate": 5e-05,
      "loss": 0.3865,
      "step": 19000
    },
    {
      "epoch": 2.39,
      "eval_loss": 0.7468438744544983,
      "eval_runtime": 4.0121,
      "eval_samples_per_second": 249.249,
      "eval_steps_per_second": 31.156,
      "step": 19000
    },
    {
      "epoch": 2.46,
      "learning_rate": 2.5e-05,
      "loss": 0.3633,
      "step": 19500
    },
    {
      "epoch": 2.46,
      "eval_loss": 0.7479762434959412,
      "eval_runtime": 4.0316,
      "eval_samples_per_second": 248.04,
      "eval_steps_per_second": 31.005,
      "step": 19500
    },
    {
      "epoch": 2.52,
      "learning_rate": 0.0,
      "loss": 0.3674,
      "step": 20000
    },
    {
      "epoch": 2.52,
      "eval_loss": 0.7462155222892761,
      "eval_runtime": 4.0244,
      "eval_samples_per_second": 248.485,
      "eval_steps_per_second": 31.061,
      "step": 20000
    }
  ],
  "logging_steps": 500,
  "max_steps": 20000,
  "num_train_epochs": 3,
  "save_steps": 5000,
  "total_flos": 3.549121832463667e+16,
  "trial_name": null,
  "trial_params": null
}