trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.121212121212121,
  "eval_steps": 500,
  "global_step": 3,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.12,
      "learning_rate": 0.00029871672920607153,
      "loss": 3.1318,
      "step": 1
    },
    {
      "epoch": 0.12,
      "eval_loss": 2.8355371952056885,
      "eval_runtime": 1.8669,
      "eval_samples_per_second": 125.339,
      "eval_steps_per_second": 4.285,
      "step": 1
    },
    {
      "epoch": 1.12,
      "eval_loss": 2.63714861869812,
      "eval_runtime": 1.5412,
      "eval_samples_per_second": 151.829,
      "eval_steps_per_second": 5.191,
      "step": 2
    },
    {
      "epoch": 2.12,
      "eval_loss": 2.496318817138672,
      "eval_runtime": 1.5238,
      "eval_samples_per_second": 153.559,
      "eval_steps_per_second": 5.25,
      "step": 3
    },
    {
      "epoch": 2.12,
      "step": 3,
      "total_flos": 270826947477504.0,
      "train_loss": 2.9835983912150064,
      "train_runtime": 146.3783,
      "train_samples_per_second": 42.978,
      "train_steps_per_second": 0.164
    }
  ],
  "logging_steps": 5,
  "max_steps": 24,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 270826947477504.0,
  "trial_name": null,
  "trial_params": null
}