trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.08683853459972862,
  "eval_steps": 500,
  "global_step": 1,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.09,
      "learning_rate": 1.9594929736144978e-05,
      "loss": 1.7095,
      "step": 1
    },
    {
      "epoch": 0.09,
      "eval_loss": 1.7014471292495728,
      "eval_runtime": 6.3775,
      "eval_samples_per_second": 46.1,
      "eval_steps_per_second": 5.802,
      "step": 1
    },
    {
      "epoch": 0.09,
      "step": 1,
      "total_flos": 5.309162136482611e+16,
      "train_loss": 2.015326350927353,
      "train_runtime": 412.8764,
      "train_samples_per_second": 14.278,
      "train_steps_per_second": 0.027
    }
  ],
  "logging_steps": 5,
  "max_steps": 11,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 5.309162136482611e+16,
  "trial_name": null,
  "trial_params": null
}