trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 1969,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.25,
      "learning_rate": 1.940180586907449e-05,
      "loss": 0.5026,
      "step": 500
    },
    {
      "epoch": 0.51,
      "learning_rate": 1.658013544018059e-05,
      "loss": 0.3654,
      "step": 1000
    },
    {
      "epoch": 0.76,
      "learning_rate": 1.3758465011286683e-05,
      "loss": 0.3256,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.24465061724185944,
      "eval_runtime": 257.734,
      "eval_samples_per_second": 14.744,
      "eval_steps_per_second": 3.686,
      "step": 1969
    }
  ],
  "logging_steps": 500,
  "max_steps": 3938,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 1.0969654814267904e+16,
  "trial_name": null,
  "trial_params": null
}