trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.5555555555555554,
  "eval_steps": 25,
  "global_step": 4,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 9.884089469909668,
      "learning_rate": 0.00015000000000000001,
      "loss": 9.5546,
      "step": 1
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 8.386604309082031,
      "learning_rate": 0.0001,
      "loss": 9.3703,
      "step": 2
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 9.898933410644531,
      "learning_rate": 5e-05,
      "loss": 9.2303,
      "step": 3
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 10.148322105407715,
      "learning_rate": 0.0,
      "loss": 9.0802,
      "step": 4
    },
    {
      "epoch": 3.5555555555555554,
      "step": 4,
      "total_flos": 3298197865788.0,
      "train_loss": 9.308886766433716,
      "train_runtime": 46.4172,
      "train_samples_per_second": 2.844,
      "train_steps_per_second": 0.086
    }
  ],
  "logging_steps": 1.0,
  "max_steps": 4,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3298197865788.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}