trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.00019238168526356292,
  "global_step": 1,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 2e-05,
      "loss": 0.5618,
      "step": 1
    },
    {
      "epoch": 0.0,
      "step": 1,
      "total_flos": 1508932845568.0,
      "train_loss": 0.56182861328125,
      "train_runtime": 31.3904,
      "train_samples_per_second": 4.078,
      "train_steps_per_second": 0.032
    }
  ],
  "max_steps": 1,
  "num_train_epochs": 1,
  "total_flos": 1508932845568.0,
  "trial_name": null,
  "trial_params": null
}