trainer_state.json

{
  "best_metric": 4.371549606323242,
  "best_model_checkpoint": "output/burzum/checkpoint-11",
  "epoch": 1.0,
  "global_step": 11,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.45,
      "learning_rate": 7.836279790554737e-05,
      "loss": 4.3563,
      "step": 5
    },
    {
      "epoch": 0.91,
      "learning_rate": 2.778782010045488e-06,
      "loss": 4.21,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_loss": 4.371549606323242,
      "eval_runtime": 0.4105,
      "eval_samples_per_second": 21.923,
      "eval_steps_per_second": 4.872,
      "step": 11
    }
  ],
  "max_steps": 11,
  "num_train_epochs": 1,
  "total_flos": 10712973312000.0,
  "trial_name": null,
  "trial_params": null
}