trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 1149,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.5,
      "learning_rate": 1.4979493526925553e-05,
      "loss": 0.2597,
      "step": 575
    },
    {
      "epoch": 1.5,
      "eval_f1": 0.9049180327868853,
      "eval_loss": 0.3436029255390167,
      "eval_runtime": 6.4623,
      "eval_samples_per_second": 210.607,
      "eval_steps_per_second": 6.654,
      "step": 575
    },
    {
      "epoch": 3.0,
      "step": 1149,
      "total_flos": 9672488617144320.0,
      "train_loss": 0.18225979452241078,
      "train_runtime": 460.3539,
      "train_samples_per_second": 79.856,
      "train_steps_per_second": 2.496
    }
  ],
  "max_steps": 1149,
  "num_train_epochs": 3,
  "total_flos": 9672488617144320.0,
  "trial_name": null,
  "trial_params": null
}