trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.6666666666666665,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.15,
      "grad_norm": 1.4186153874302803,
      "learning_rate": 1e-05,
      "loss": 1.1959,
      "step": 1
    },
    {
      "epoch": 0.74,
      "grad_norm": 2.7709253330133508,
      "learning_rate": 1.8314696123025456e-05,
      "loss": 1.1778,
      "step": 5
    },
    {
      "epoch": 0.89,
      "eval_loss": 1.0390011072158813,
      "eval_runtime": 31.8949,
      "eval_samples_per_second": 11.475,
      "eval_steps_per_second": 0.721,
      "step": 6
    },
    {
      "epoch": 1.48,
      "grad_norm": 0.840647966467128,
      "learning_rate": 1e-05,
      "loss": 1.0295,
      "step": 10
    },
    {
      "epoch": 1.93,
      "eval_loss": 0.8658971190452576,
      "eval_runtime": 31.8683,
      "eval_samples_per_second": 11.485,
      "eval_steps_per_second": 0.722,
      "step": 13
    },
    {
      "epoch": 2.22,
      "grad_norm": 0.685152903494503,
      "learning_rate": 1.6853038769745466e-06,
      "loss": 0.903,
      "step": 15
    },
    {
      "epoch": 2.67,
      "eval_loss": 0.8405471444129944,
      "eval_runtime": 32.0919,
      "eval_samples_per_second": 11.405,
      "eval_steps_per_second": 0.717,
      "step": 18
    },
    {
      "epoch": 2.67,
      "step": 18,
      "total_flos": 3716488888320.0,
      "train_loss": 1.002972490257687,
      "train_runtime": 396.6798,
      "train_samples_per_second": 3.222,
      "train_steps_per_second": 0.045
    }
  ],
  "logging_steps": 5,
  "max_steps": 18,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3716488888320.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}