trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1541,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.65,
      "grad_norm": 1.8373916149139404,
      "learning_rate": 1.3816303042608856e-05,
      "loss": 0.7092,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "step": 1541,
      "total_flos": 1.1046022865525146e+17,
      "train_loss": 0.7010668161084016,
      "train_runtime": 7630.7473,
      "train_samples_per_second": 0.808,
      "train_steps_per_second": 0.202
    }
  ],
  "logging_steps": 1000,
  "max_steps": 1541,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 1000,
  "total_flos": 1.1046022865525146e+17,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}