trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9994761655316919,
  "global_step": 477,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1.9958071278825996e-05,
      "loss": 2.9192,
      "step": 1
    },
    {
      "epoch": 1.0,
      "step": 477,
      "total_flos": 1.482031706973143e+17,
      "train_loss": 2.389765337578156,
      "train_runtime": 86824.1392,
      "train_samples_per_second": 0.044,
      "train_steps_per_second": 0.005
    }
  ],
  "max_steps": 477,
  "num_train_epochs": 1,
  "total_flos": 1.482031706973143e+17,
  "trial_name": null,
  "trial_params": null
}