{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.6989911228127402,
  "global_step": 30000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 5.999999999999999e-06,
      "loss": 0.9072,
      "step": 500
    },
    {
      "epoch": 0.02,
      "learning_rate": 1.1999999999999999e-05,
      "loss": 0.6779,
      "step": 1000
    },
    {
      "epoch": 0.02,
      "eval_loss": 0.6512336730957031,
      "eval_runtime": 1.3899,
      "eval_samples_per_second": 1581.436,
      "eval_steps_per_second": 25.182,
      "step": 1000
    },
    {
      "epoch": 0.03,
      "learning_rate": 1.7999999999999997e-05,
      "loss": 0.6516,
      "step": 1500
    },
    {
      "epoch": 0.05,
      "learning_rate": 2.3999999999999997e-05,
      "loss": 0.6505,
      "step": 2000
    },
    {
      "epoch": 0.05,
      "eval_loss": 0.6479398608207703,
      "eval_runtime": 1.2819,
      "eval_samples_per_second": 1714.632,
      "eval_steps_per_second": 27.303,
      "step": 2000
    },
    {
      "epoch": 0.06,
      "learning_rate": 2.9999999999999997e-05,
      "loss": 0.6487,
      "step": 2500
    },
    {
      "epoch": 0.07,
      "learning_rate": 3.5999999999999994e-05,
      "loss": 0.6475,
      "step": 3000
    },
    {
      "epoch": 0.07,
      "eval_loss": 0.6461147665977478,
      "eval_runtime": 1.2932,
      "eval_samples_per_second": 1699.633,
      "eval_steps_per_second": 27.064,
      "step": 3000
    },
    {
      "epoch": 0.08,
      "learning_rate": 4.2e-05,
      "loss": 0.6469,
      "step": 3500
    },
    {
      "epoch": 0.09,
      "learning_rate": 4.7999999999999994e-05,
      "loss": 0.6464,
      "step": 4000
    },
    {
      "epoch": 0.09,
      "eval_loss": 0.6451290249824524,
      "eval_runtime": 1.2355,
      "eval_samples_per_second": 1779.085,
      "eval_steps_per_second": 28.329,
      "step": 4000
    },
    {
      "epoch": 0.1,
      "learning_rate": 5.399999999999999e-05,
      "loss": 0.6459,
      "step": 4500
    },
    {
      "epoch": 0.12,
      "learning_rate": 5.9999999999999995e-05,
      "loss": 0.6455,
      "step": 5000
    },
    {
      "epoch": 0.12,
      "eval_loss": 0.6444417238235474,
      "eval_runtime": 1.2972,
      "eval_samples_per_second": 1694.435,
      "eval_steps_per_second": 26.981,
      "step": 5000
    },
    {
      "epoch": 0.13,
      "learning_rate": 6.599999999999999e-05,
      "loss": 0.6452,
      "step": 5500
    },
    {
      "epoch": 0.14,
      "learning_rate": 7.199999999999999e-05,
      "loss": 0.6448,
      "step": 6000
    },
    {
      "epoch": 0.14,
      "eval_loss": 0.6438854336738586,
      "eval_runtime": 1.2871,
      "eval_samples_per_second": 1707.767,
      "eval_steps_per_second": 27.194,
      "step": 6000
    },
    {
      "epoch": 0.15,
      "learning_rate": 7.8e-05,
      "loss": 0.64,
      "step": 6500
    },
    {
      "epoch": 0.16,
      "learning_rate": 8.4e-05,
      "loss": 0.6222,
      "step": 7000
    },
    {
      "epoch": 0.16,
      "eval_loss": 0.6123253703117371,
      "eval_runtime": 1.2501,
      "eval_samples_per_second": 1758.21,
      "eval_steps_per_second": 27.997,
      "step": 7000
    },
    {
      "epoch": 0.17,
      "learning_rate": 8.999999999999999e-05,
      "loss": 0.6115,
      "step": 7500
    },
    {
      "epoch": 0.19,
      "learning_rate": 9.599999999999999e-05,
      "loss": 0.6019,
      "step": 8000
    },
    {
      "epoch": 0.19,
      "eval_loss": 0.5872684717178345,
      "eval_runtime": 1.2778,
      "eval_samples_per_second": 1720.181,
      "eval_steps_per_second": 27.391,
      "step": 8000
    },
    {
      "epoch": 0.2,
      "learning_rate": 0.000102,
      "loss": 0.5924,
      "step": 8500
    },
    {
      "epoch": 0.21,
      "learning_rate": 0.00010799999999999998,
      "loss": 0.5838,
      "step": 9000
    },
    {
      "epoch": 0.21,
      "eval_loss": 0.5689075589179993,
      "eval_runtime": 1.2743,
      "eval_samples_per_second": 1724.838,
      "eval_steps_per_second": 27.466,
      "step": 9000
    },
    {
      "epoch": 0.22,
      "learning_rate": 0.00011399999999999999,
      "loss": 0.5737,
      "step": 9500
    },
    {
      "epoch": 0.23,
      "learning_rate": 0.00011999999999999999,
      "loss": 0.5612,
      "step": 10000
    },
    {
      "epoch": 0.23,
      "eval_loss": 0.5473007559776306,
      "eval_runtime": 1.2648,
      "eval_samples_per_second": 1737.882,
      "eval_steps_per_second": 27.673,
      "step": 10000
    },
    {
      "epoch": 0.24,
      "learning_rate": 0.00012599999999999997,
      "loss": 0.5516,
      "step": 10500
    },
    {
      "epoch": 0.26,
      "learning_rate": 0.00013199999999999998,
      "loss": 0.5428,
      "step": 11000
    },
    {
      "epoch": 0.26,
      "eval_loss": 0.5300613045692444,
      "eval_runtime": 1.2658,
      "eval_samples_per_second": 1736.485,
      "eval_steps_per_second": 27.651,
      "step": 11000
    },
    {
      "epoch": 0.27,
      "learning_rate": 0.000138,
      "loss": 0.5346,
      "step": 11500
    },
    {
      "epoch": 0.28,
      "learning_rate": 0.00014399999999999998,
      "loss": 0.527,
      "step": 12000
    },
    {
      "epoch": 0.28,
      "eval_loss": 0.5140743255615234,
      "eval_runtime": 1.2609,
      "eval_samples_per_second": 1743.192,
      "eval_steps_per_second": 27.758,
      "step": 12000
    },
    {
      "epoch": 0.29,
      "learning_rate": 0.00015,
      "loss": 0.5203,
      "step": 12500
    },
    {
      "epoch": 0.3,
      "learning_rate": 0.000156,
      "loss": 0.5142,
      "step": 13000
    },
    {
      "epoch": 0.3,
      "eval_loss": 0.5022566318511963,
      "eval_runtime": 1.2481,
      "eval_samples_per_second": 1761.066,
      "eval_steps_per_second": 28.042,
      "step": 13000
    },
    {
      "epoch": 0.31,
      "learning_rate": 0.000162,
      "loss": 0.5079,
      "step": 13500
    },
    {
      "epoch": 0.33,
      "learning_rate": 0.000168,
      "loss": 0.5002,
      "step": 14000
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.4841987192630768,
      "eval_runtime": 1.3176,
      "eval_samples_per_second": 1668.178,
      "eval_steps_per_second": 26.563,
      "step": 14000
    },
    {
      "epoch": 0.34,
      "learning_rate": 0.00017399999999999997,
      "loss": 0.4916,
      "step": 14500
    },
    {
      "epoch": 0.35,
      "learning_rate": 0.00017999999999999998,
      "loss": 0.4852,
      "step": 15000
    },
    {
      "epoch": 0.35,
      "eval_loss": 0.47066375613212585,
      "eval_runtime": 1.3026,
      "eval_samples_per_second": 1687.422,
      "eval_steps_per_second": 26.87,
      "step": 15000
    },
    {
      "epoch": 0.36,
      "learning_rate": 0.000186,
      "loss": 0.4778,
      "step": 15500
    },
    {
      "epoch": 0.37,
      "learning_rate": 0.00019199999999999998,
      "loss": 0.4697,
      "step": 16000
    },
    {
      "epoch": 0.37,
      "eval_loss": 0.4562186300754547,
      "eval_runtime": 1.2895,
      "eval_samples_per_second": 1704.503,
      "eval_steps_per_second": 27.142,
      "step": 16000
    },
    {
      "epoch": 0.38,
      "learning_rate": 0.000198,
      "loss": 0.4636,
      "step": 16500
    },
    {
      "epoch": 0.4,
      "learning_rate": 0.000204,
      "loss": 0.459,
      "step": 17000
    },
    {
      "epoch": 0.4,
      "eval_loss": 0.44295796751976013,
      "eval_runtime": 1.2516,
      "eval_samples_per_second": 1756.093,
      "eval_steps_per_second": 27.963,
      "step": 17000
    },
    {
      "epoch": 0.41,
      "learning_rate": 0.00020999999999999998,
      "loss": 0.4544,
      "step": 17500
    },
    {
      "epoch": 0.42,
      "learning_rate": 0.00021599999999999996,
      "loss": 0.45,
      "step": 18000
    },
    {
      "epoch": 0.42,
      "eval_loss": 0.4343169629573822,
      "eval_runtime": 1.327,
      "eval_samples_per_second": 1656.351,
      "eval_steps_per_second": 26.375,
      "step": 18000
    },
    {
      "epoch": 0.43,
      "learning_rate": 0.00022199999999999998,
      "loss": 0.4458,
      "step": 18500
    },
    {
      "epoch": 0.44,
      "learning_rate": 0.00022799999999999999,
      "loss": 0.4411,
      "step": 19000
    },
    {
      "epoch": 0.44,
      "eval_loss": 0.42496559023857117,
      "eval_runtime": 1.3285,
      "eval_samples_per_second": 1654.492,
      "eval_steps_per_second": 26.345,
      "step": 19000
    },
    {
      "epoch": 0.45,
      "learning_rate": 0.000234,
      "loss": 0.4368,
      "step": 19500
    },
    {
      "epoch": 0.47,
      "learning_rate": 0.00023999999999999998,
      "loss": 0.4322,
      "step": 20000
    },
    {
      "epoch": 0.47,
      "eval_loss": 0.4132905900478363,
      "eval_runtime": 1.3158,
      "eval_samples_per_second": 1670.425,
      "eval_steps_per_second": 26.599,
      "step": 20000
    },
    {
      "epoch": 0.48,
      "learning_rate": 0.00024599999999999996,
      "loss": 0.4278,
      "step": 20500
    },
    {
      "epoch": 0.49,
      "learning_rate": 0.00025199999999999995,
      "loss": 0.4233,
      "step": 21000
    },
    {
      "epoch": 0.49,
      "eval_loss": 0.4032529294490814,
      "eval_runtime": 1.2842,
      "eval_samples_per_second": 1711.577,
      "eval_steps_per_second": 27.254,
      "step": 21000
    },
    {
      "epoch": 0.5,
      "learning_rate": 0.000258,
      "loss": 0.4189,
      "step": 21500
    },
    {
      "epoch": 0.51,
      "learning_rate": 0.00026399999999999997,
      "loss": 0.4146,
      "step": 22000
    },
    {
      "epoch": 0.51,
      "eval_loss": 0.39436739683151245,
      "eval_runtime": 1.2633,
      "eval_samples_per_second": 1739.901,
      "eval_steps_per_second": 27.705,
      "step": 22000
    },
    {
      "epoch": 0.52,
      "learning_rate": 0.00027,
      "loss": 0.4109,
      "step": 22500
    },
    {
      "epoch": 0.54,
      "learning_rate": 0.000276,
      "loss": 0.4077,
      "step": 23000
    },
    {
      "epoch": 0.54,
      "eval_loss": 0.38647642731666565,
      "eval_runtime": 1.2503,
      "eval_samples_per_second": 1758.021,
      "eval_steps_per_second": 27.994,
      "step": 23000
    },
    {
      "epoch": 0.55,
      "learning_rate": 0.00028199999999999997,
      "loss": 0.4047,
      "step": 23500
    },
    {
      "epoch": 0.56,
      "learning_rate": 0.00028799999999999995,
      "loss": 0.4019,
      "step": 24000
    },
    {
      "epoch": 0.56,
      "eval_loss": 0.381785124540329,
      "eval_runtime": 1.2485,
      "eval_samples_per_second": 1760.518,
      "eval_steps_per_second": 28.034,
      "step": 24000
    },
    {
      "epoch": 0.57,
      "learning_rate": 0.000294,
      "loss": 0.3994,
      "step": 24500
    },
    {
      "epoch": 0.58,
      "learning_rate": 0.0003,
      "loss": 0.3968,
      "step": 25000
    },
    {
      "epoch": 0.58,
      "eval_loss": 0.3785216510295868,
      "eval_runtime": 1.2913,
      "eval_samples_per_second": 1702.153,
      "eval_steps_per_second": 27.104,
      "step": 25000
    },
    {
      "epoch": 0.59,
      "learning_rate": 0.00029999920715161553,
      "loss": 0.3945,
      "step": 25500
    },
    {
      "epoch": 0.61,
      "learning_rate": 0.0002999968286151326,
      "loss": 0.392,
      "step": 26000
    },
    {
      "epoch": 0.61,
      "eval_loss": 0.37318962812423706,
      "eval_runtime": 1.2758,
      "eval_samples_per_second": 1722.782,
      "eval_steps_per_second": 27.433,
      "step": 26000
    },
    {
      "epoch": 0.62,
      "learning_rate": 0.0002999928644165624,
      "loss": 0.3897,
      "step": 26500
    },
    {
      "epoch": 0.63,
      "learning_rate": 0.0002999873145992569,
      "loss": 0.3869,
      "step": 27000
    },
    {
      "epoch": 0.63,
      "eval_loss": 0.3669659197330475,
      "eval_runtime": 1.2869,
      "eval_samples_per_second": 1707.967,
      "eval_steps_per_second": 27.197,
      "step": 27000
    },
    {
      "epoch": 0.64,
      "learning_rate": 0.000299980179223908,
      "loss": 0.385,
      "step": 27500
    },
    {
      "epoch": 0.65,
      "learning_rate": 0.0002999714583685469,
      "loss": 0.3828,
      "step": 28000
    },
    {
      "epoch": 0.65,
      "eval_loss": 0.3615981340408325,
      "eval_runtime": 1.2961,
      "eval_samples_per_second": 1695.914,
      "eval_steps_per_second": 27.005,
      "step": 28000
    },
    {
      "epoch": 0.66,
      "learning_rate": 0.00029996115212854366,
      "loss": 0.3808,
      "step": 28500
    },
    {
      "epoch": 0.68,
      "learning_rate": 0.00029994926061660554,
      "loss": 0.3786,
      "step": 29000
    },
    {
      "epoch": 0.68,
      "eval_loss": 0.3613170087337494,
      "eval_runtime": 1.3013,
      "eval_samples_per_second": 1689.114,
      "eval_steps_per_second": 26.897,
      "step": 29000
    },
    {
      "epoch": 0.69,
      "learning_rate": 0.0002999357839627762,
      "loss": 0.377,
      "step": 29500
    },
    {
      "epoch": 0.7,
      "learning_rate": 0.00029992072231443425,
      "loss": 0.3751,
      "step": 30000
    },
    {
      "epoch": 0.7,
      "eval_loss": 0.35551005601882935,
      "eval_runtime": 1.2661,
      "eval_samples_per_second": 1736.053,
      "eval_steps_per_second": 27.644,
      "step": 30000
    }
  ],
  "max_steps": 500000,
  "num_train_epochs": 12,
  "total_flos": 9.584614163235e+20,
  "trial_name": null,
  "trial_params": null
}