trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.512,
  "eval_steps": 500,
  "global_step": 1,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.51,
      "learning_rate": 0.0,
      "loss": 1.1217,
      "step": 1
    },
    {
      "epoch": 0.51,
      "eval_loss": 1.1351821422576904,
      "eval_runtime": 5862.1633,
      "eval_samples_per_second": 3.942,
      "eval_steps_per_second": 0.493,
      "step": 1
    },
    {
      "epoch": 0.51,
      "step": 1,
      "total_flos": 4.507945275280589e+16,
      "train_loss": 1.1217185258865356,
      "train_runtime": 6493.465,
      "train_samples_per_second": 0.154,
      "train_steps_per_second": 0.0
    }
  ],
  "logging_steps": 5,
  "max_steps": 1,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 4.507945275280589e+16,
  "trial_name": null,
  "trial_params": null
}