trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "global_step": 620,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 8.06,
      "learning_rate": 9.67741935483871e-06,
      "loss": 3.5199,
      "step": 500
    },
    {
      "epoch": 10.0,
      "step": 620,
      "total_flos": 1.367278994866176e+16,
      "train_loss": 3.161412368282195,
      "train_runtime": 1377.1007,
      "train_samples_per_second": 22.497,
      "train_steps_per_second": 0.45
    }
  ],
  "max_steps": 620,
  "num_train_epochs": 10,
  "total_flos": 1.367278994866176e+16,
  "trial_name": null,
  "trial_params": null
}