trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.333333333333333,
  "eval_steps": 500,
  "global_step": 8,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 1.9667768478393555,
      "learning_rate": 0.00015000000000000001,
      "loss": 9.3086,
      "step": 2
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 1.3830890655517578,
      "learning_rate": 0.0001,
      "loss": 7.9088,
      "step": 4
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.0656214952468872,
      "learning_rate": 5e-05,
      "loss": 7.4385,
      "step": 6
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 0.9365198016166687,
      "learning_rate": 0.0,
      "loss": 7.2958,
      "step": 8
    },
    {
      "epoch": 5.333333333333333,
      "step": 8,
      "total_flos": 36142266989280.0,
      "train_loss": 7.987936019897461,
      "train_runtime": 213.0103,
      "train_samples_per_second": 0.901,
      "train_steps_per_second": 0.038
    }
  ],
  "logging_steps": 2,
  "max_steps": 8,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": false,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 36142266989280.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}