trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.23923444976076555,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.24,
      "learning_rate": 1e-07,
      "loss": 0.0164,
      "step": 100
    },
    {
      "epoch": 0.24,
      "eval_accuracy_sunlit": NaN,
      "eval_iou_sunlit": 0.0,
      "eval_loss": 0.01966829039156437,
      "eval_mean_accuracy": NaN,
      "eval_mean_iou": 0.0,
      "eval_overall_accuracy": NaN,
      "eval_runtime": 1173.2723,
      "eval_samples_per_second": 0.936,
      "eval_steps_per_second": 0.118,
      "step": 100
    },
    {
      "epoch": 0.24,
      "step": 100,
      "total_flos": 1.40213942747136e+16,
      "train_loss": 0.016446611881256103,
      "train_runtime": 1220.807,
      "train_samples_per_second": 0.655,
      "train_steps_per_second": 0.082
    }
  ],
  "logging_steps": 100,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 1.40213942747136e+16,
  "trial_name": null,
  "trial_params": null
}