trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9998933560840354,
  "global_step": 2344,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.21,
      "learning_rate": 1.7486960644855383e-05,
      "loss": 1.4078,
      "step": 500
    },
    {
      "epoch": 0.43,
      "learning_rate": 1.2745376955903272e-05,
      "loss": 1.1637,
      "step": 1000
    },
    {
      "epoch": 0.64,
      "learning_rate": 8.003793266951162e-06,
      "loss": 1.0799,
      "step": 1500
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.262209577999052e-06,
      "loss": 1.0642,
      "step": 2000
    },
    {
      "epoch": 1.0,
      "eval_loss": 1.197797417640686,
      "eval_runtime": 2.2647,
      "eval_samples_per_second": 137.328,
      "step": 2344
    }
  ],
  "max_steps": 2344,
  "num_train_epochs": 1,
  "total_flos": 5734497329676288.0,
  "trial_name": null,
  "trial_params": null
}