{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 276,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "learning_rate": 1.992753623188406e-05,
      "loss": 2.7608,
      "step": 1
    },
    {
      "epoch": 3.0,
      "step": 276,
      "total_flos": 1.637356822069248e+16,
      "train_loss": 1.5195069511731465,
      "train_runtime": 1160.0325,
      "train_samples_per_second": 1.901,
      "train_steps_per_second": 0.238
    }
  ],
  "max_steps": 276,
  "num_train_epochs": 3,
  "total_flos": 1.637356822069248e+16,
  "trial_name": null,
  "trial_params": null
}