trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 27,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 0.6436542272567749,
      "eval_runtime": 0.5454,
      "eval_samples_per_second": 33.001,
      "eval_steps_per_second": 5.5,
      "step": 9
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.6364022493362427,
      "eval_runtime": 0.5564,
      "eval_samples_per_second": 32.353,
      "eval_steps_per_second": 5.392,
      "step": 18
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.6323569416999817,
      "eval_runtime": 0.56,
      "eval_samples_per_second": 32.145,
      "eval_steps_per_second": 5.357,
      "step": 27
    },
    {
      "epoch": 3.0,
      "step": 27,
      "total_flos": 56042654791680.0,
      "train_loss": 0.6852263697871456,
      "train_runtime": 23.2921,
      "train_samples_per_second": 9.145,
      "train_steps_per_second": 1.159
    }
  ],
  "logging_steps": 500,
  "max_steps": 27,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 56042654791680.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}