checkpoint-16838/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "global_step": 16838,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.394,
      "step": 16838
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8967889908256881,
      "eval_loss": 0.3896108567714691,
      "eval_runtime": 19.0439,
      "eval_samples_per_second": 45.789,
      "eval_steps_per_second": 5.724,
      "step": 16838
    }
  ],
  "max_steps": 50514,
  "num_train_epochs": 3,
  "total_flos": 4.213461867364147e+16,
  "trial_name": null,
  "trial_params": null
}