trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 891,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 2.170140027999878,
      "eval_runtime": 15.6956,
      "eval_samples_per_second": 63.712,
      "eval_steps_per_second": 7.964,
      "step": 297
    },
    {
      "epoch": 1.68,
      "learning_rate": 8.77665544332211e-06,
      "loss": 2.1401,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_loss": 2.0152316093444824,
      "eval_runtime": 15.6747,
      "eval_samples_per_second": 63.797,
      "eval_steps_per_second": 7.975,
      "step": 594
    },
    {
      "epoch": 3.0,
      "eval_loss": 2.1160194873809814,
      "eval_runtime": 15.6683,
      "eval_samples_per_second": 63.823,
      "eval_steps_per_second": 7.978,
      "step": 891
    },
    {
      "epoch": 3.0,
      "step": 891,
      "total_flos": 1.5005754149105664e+16,
      "train_loss": 2.1000711460306185,
      "train_runtime": 3427.7755,
      "train_samples_per_second": 16.632,
      "train_steps_per_second": 0.26
    }
  ],
  "max_steps": 891,
  "num_train_epochs": 3,
  "total_flos": 1.5005754149105664e+16,
  "trial_name": null,
  "trial_params": null
}