trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 9.99442275515895,
  "eval_steps": 500,
  "global_step": 4480,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.12,
      "grad_norm": 18.684795379638672,
      "learning_rate": 4.4419642857142854e-05,
      "loss": 0.6311,
      "step": 500
    },
    {
      "epoch": 2.23,
      "grad_norm": 13.164822578430176,
      "learning_rate": 3.883928571428572e-05,
      "loss": 0.3308,
      "step": 1000
    },
    {
      "epoch": 3.35,
      "grad_norm": 17.804838180541992,
      "learning_rate": 3.325892857142857e-05,
      "loss": 0.209,
      "step": 1500
    },
    {
      "epoch": 4.46,
      "grad_norm": 4.987255096435547,
      "learning_rate": 2.767857142857143e-05,
      "loss": 0.1576,
      "step": 2000
    },
    {
      "epoch": 5.58,
      "grad_norm": 9.00122356414795,
      "learning_rate": 2.2098214285714286e-05,
      "loss": 0.1173,
      "step": 2500
    },
    {
      "epoch": 6.69,
      "grad_norm": 1.4200397729873657,
      "learning_rate": 1.6517857142857144e-05,
      "loss": 0.082,
      "step": 3000
    },
    {
      "epoch": 7.81,
      "grad_norm": 2.2957751750946045,
      "learning_rate": 1.09375e-05,
      "loss": 0.061,
      "step": 3500
    },
    {
      "epoch": 8.92,
      "grad_norm": 0.13178028166294098,
      "learning_rate": 5.357142857142857e-06,
      "loss": 0.0428,
      "step": 4000
    },
    {
      "epoch": 9.99,
      "step": 4480,
      "total_flos": 2.577776514283278e+16,
      "train_loss": 0.1851580994469779,
      "train_runtime": 8550.1342,
      "train_samples_per_second": 16.769,
      "train_steps_per_second": 0.524
    }
  ],
  "logging_steps": 500,
  "max_steps": 4480,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 2.577776514283278e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}