trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.5555555555555554,
  "eval_steps": 500,
  "global_step": 12,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5925925925925926,
      "grad_norm": 9.149438858032227,
      "learning_rate": 0.0001666666666666667,
      "loss": 75.8026,
      "step": 2
    },
    {
      "epoch": 1.1851851851851851,
      "grad_norm": 10.327601432800293,
      "learning_rate": 0.00013333333333333334,
      "loss": 72.9765,
      "step": 4
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 16.77288818359375,
      "learning_rate": 0.0001,
      "loss": 69.6137,
      "step": 6
    },
    {
      "epoch": 2.3703703703703702,
      "grad_norm": NaN,
      "learning_rate": 8.333333333333334e-05,
      "loss": 65.7295,
      "step": 8
    },
    {
      "epoch": 2.962962962962963,
      "grad_norm": 14.90024185180664,
      "learning_rate": 5e-05,
      "loss": 63.965,
      "step": 10
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 11.715697288513184,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 62.6076,
      "step": 12
    },
    {
      "epoch": 3.5555555555555554,
      "step": 12,
      "total_flos": 59016492903792.0,
      "train_loss": 68.44916152954102,
      "train_runtime": 340.8103,
      "train_samples_per_second": 0.634,
      "train_steps_per_second": 0.035
    }
  ],
  "logging_steps": 2,
  "max_steps": 12,
  "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": 59016492903792.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}