trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9978245105148659,
  "global_step": 344,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.29,
      "learning_rate": 3.7125e-06,
      "loss": 12.2666,
      "step": 100
    },
    {
      "epoch": 0.58,
      "learning_rate": 7.4625e-06,
      "loss": 6.1937,
      "step": 200
    },
    {
      "epoch": 0.87,
      "learning_rate": 1.1212499999999998e-05,
      "loss": 4.1366,
      "step": 300
    },
    {
      "epoch": 1.0,
      "step": 344,
      "total_flos": 1.0141067319340954e+18,
      "train_loss": 7.050228030182595,
      "train_runtime": 437.537,
      "train_samples_per_second": 25.209,
      "train_steps_per_second": 0.786
    }
  ],
  "max_steps": 344,
  "num_train_epochs": 1,
  "total_flos": 1.0141067319340954e+18,
  "trial_name": null,
  "trial_params": null
}