trainer_state.json

{
  "best_global_step": 21000,
  "best_metric": 0.9027300735095976,
  "best_model_checkpoint": "./saved_best/sst2_gpt2-20000/full_text/checkpoint-21000",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 21000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.07762611657381058,
      "learning_rate": 1.3334285714285714e-05,
      "loss": 2.1968,
      "step": 7000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8826666666666667,
      "eval_f1_macro": 0.8820264504294919,
      "eval_f1_weighted": 0.8829302851172682,
      "eval_loss": 1.7135562896728516,
      "eval_runtime": 270.0093,
      "eval_samples_per_second": 22.221,
      "eval_steps_per_second": 11.111,
      "step": 7000
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.649629533290863,
      "learning_rate": 6.667619047619049e-06,
      "loss": 1.5723,
      "step": 14000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8971666666666667,
      "eval_f1_macro": 0.8958877488819252,
      "eval_f1_weighted": 0.8970878160964796,
      "eval_loss": 1.7208778858184814,
      "eval_runtime": 270.5601,
      "eval_samples_per_second": 22.176,
      "eval_steps_per_second": 11.088,
      "step": 14000
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.15791520476341248,
      "learning_rate": 9.523809523809524e-10,
      "loss": 1.0602,
      "step": 21000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9038333333333334,
      "eval_f1_macro": 0.9027300735095976,
      "eval_f1_weighted": 0.9038074352154053,
      "eval_loss": 1.6445306539535522,
      "eval_runtime": 271.1742,
      "eval_samples_per_second": 22.126,
      "eval_steps_per_second": 11.063,
      "step": 21000
    },
    {
      "epoch": 3.0,
      "step": 21000,
      "total_flos": 1328938942464000.0,
      "train_loss": 1.609793224516369,
      "train_runtime": 4749.0538,
      "train_samples_per_second": 8.844,
      "train_steps_per_second": 4.422
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9038333333333334,
      "eval_f1_macro": 0.9027300735095976,
      "eval_f1_weighted": 0.9038074352154053,
      "eval_loss": 1.6445306539535522,
      "eval_runtime": 270.933,
      "eval_samples_per_second": 22.146,
      "eval_steps_per_second": 11.073,
      "step": 21000
    }
  ],
  "logging_steps": 10,
  "max_steps": 21000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 1328938942464000.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}