{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 9723,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.2,
      "learning_rate": 1.997944501541624e-05,
      "loss": 1.1561,
      "step": 648
    },
    {
      "epoch": 0.4,
      "learning_rate": 2.889257142857143e-05,
      "loss": 0.9124,
      "step": 1296
    },
    {
      "epoch": 0.6,
      "learning_rate": 2.6670857142857143e-05,
      "loss": 0.8886,
      "step": 1944
    },
    {
      "epoch": 0.8,
      "learning_rate": 2.444914285714286e-05,
      "loss": 0.8963,
      "step": 2592
    },
    {
      "epoch": 1.0,
      "learning_rate": 2.2227428571428572e-05,
      "loss": 0.8916,
      "step": 3240
    },
    {
      "epoch": 1.2,
      "learning_rate": 2.0005714285714285e-05,
      "loss": 0.8783,
      "step": 3888
    },
    {
      "epoch": 1.4,
      "learning_rate": 1.7784e-05,
      "loss": 0.8789,
      "step": 4536
    },
    {
      "epoch": 1.6,
      "learning_rate": 1.5569142857142856e-05,
      "loss": 0.8707,
      "step": 5184
    },
    {
      "epoch": 1.8,
      "learning_rate": 1.3347428571428572e-05,
      "loss": 0.8713,
      "step": 5832
    },
    {
      "epoch": 2.0,
      "learning_rate": 1.1125714285714285e-05,
      "loss": 0.8737,
      "step": 6480
    },
    {
      "epoch": 2.2,
      "learning_rate": 8.907428571428572e-06,
      "loss": 0.8639,
      "step": 7128
    },
    {
      "epoch": 2.4,
      "learning_rate": 6.685714285714286e-06,
      "loss": 0.8625,
      "step": 7776
    },
    {
      "epoch": 2.6,
      "learning_rate": 4.464e-06,
      "loss": 0.8675,
      "step": 8424
    },
    {
      "epoch": 2.8,
      "learning_rate": 2.242285714285714e-06,
      "loss": 0.8682,
      "step": 9072
    },
    {
      "epoch": 3.0,
      "learning_rate": 2.0571428571428573e-08,
      "loss": 0.8641,
      "step": 9720
    }
  ],
  "logging_steps": 648,
  "max_steps": 9723,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 7.903007639422894e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}