checkpoint-50/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.03571428571428571,
  "eval_steps": 500,
  "global_step": 50,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.014285714285714285,
      "grad_norm": 2.227086305618286,
      "learning_rate": 1.4285714285714285e-05,
      "loss": 2.7252,
      "step": 20
    },
    {
      "epoch": 0.02857142857142857,
      "grad_norm": 1.4889636039733887,
      "learning_rate": 2.857142857142857e-05,
      "loss": 2.3403,
      "step": 40
    }
  ],
  "logging_steps": 20,
  "max_steps": 1400,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 300681894297600.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}