trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.7272727272727275,
  "eval_steps": 500,
  "global_step": 15,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.18,
      "grad_norm": 8.11206379440835,
      "learning_rate": 2.5e-06,
      "loss": 1.4798,
      "step": 1
    },
    {
      "epoch": 0.91,
      "grad_norm": 4.350967687374259,
      "learning_rate": 4.3712768704277535e-06,
      "loss": 1.373,
      "step": 5
    },
    {
      "epoch": 0.91,
      "eval_loss": 1.3393189907073975,
      "eval_runtime": 1.3211,
      "eval_samples_per_second": 5.299,
      "eval_steps_per_second": 0.757,
      "step": 5
    },
    {
      "epoch": 1.82,
      "grad_norm": 2.5602166521817993,
      "learning_rate": 1.613487782393661e-06,
      "loss": 1.1772,
      "step": 10
    },
    {
      "epoch": 2.0,
      "eval_loss": 1.2193868160247803,
      "eval_runtime": 1.372,
      "eval_samples_per_second": 5.102,
      "eval_steps_per_second": 0.729,
      "step": 11
    },
    {
      "epoch": 2.73,
      "grad_norm": 1.7983907201847489,
      "learning_rate": 0.0,
      "loss": 1.0821,
      "step": 15
    },
    {
      "epoch": 2.73,
      "eval_loss": 1.2026808261871338,
      "eval_runtime": 1.3708,
      "eval_samples_per_second": 5.106,
      "eval_steps_per_second": 0.729,
      "step": 15
    },
    {
      "epoch": 2.73,
      "step": 15,
      "total_flos": 3088349921280.0,
      "train_loss": 1.2178923686345418,
      "train_runtime": 249.7212,
      "train_samples_per_second": 4.217,
      "train_steps_per_second": 0.06
    }
  ],
  "logging_steps": 5,
  "max_steps": 15,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 3088349921280.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}