trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 6.0,
  "global_step": 2244,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.34,
      "learning_rate": 3.885918003565063e-05,
      "loss": 2.1414,
      "step": 500
    },
    {
      "epoch": 2.67,
      "learning_rate": 2.7718360071301247e-05,
      "loss": 1.7805,
      "step": 1000
    },
    {
      "epoch": 4.01,
      "learning_rate": 1.6577540106951873e-05,
      "loss": 1.6037,
      "step": 1500
    },
    {
      "epoch": 5.35,
      "learning_rate": 5.436720142602496e-06,
      "loss": 1.5264,
      "step": 2000
    },
    {
      "epoch": 6.0,
      "step": 2244,
      "total_flos": 1172678639616000.0,
      "train_loss": 1.7339835345426344,
      "train_runtime": 586.4467,
      "train_samples_per_second": 3.826,
      "train_steps_per_second": 3.826
    }
  ],
  "max_steps": 2244,
  "num_train_epochs": 6,
  "total_flos": 1172678639616000.0,
  "trial_name": null,
  "trial_params": null
}