trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.0,
  "global_step": 1340,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.87,
      "learning_rate": 1.2537313432835823e-05,
      "loss": 0.41,
      "step": 500
    },
    {
      "epoch": 3.73,
      "learning_rate": 5.074626865671642e-06,
      "loss": 0.1757,
      "step": 1000
    },
    {
      "epoch": 5.0,
      "step": 1340,
      "total_flos": 2812328292979200.0,
      "train_loss": 0.24051682059444598,
      "train_runtime": 776.8272,
      "train_samples_per_second": 55.038,
      "train_steps_per_second": 1.725
    }
  ],
  "max_steps": 1340,
  "num_train_epochs": 5,
  "total_flos": 2812328292979200.0,
  "trial_name": null,
  "trial_params": null
}