run-0/checkpoint-536/trainer_state.json

{
  "best_metric": 0.43765045412808823,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-0/checkpoint-402",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 536,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_loss": 0.5353219509124756,
      "eval_matthews_correlation": 0.2835704013245166,
      "eval_runtime": 1.3472,
      "eval_samples_per_second": 774.178,
      "eval_steps_per_second": 48.989,
      "step": 134
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.514467716217041,
      "eval_matthews_correlation": 0.4210344655831915,
      "eval_runtime": 0.8237,
      "eval_samples_per_second": 1266.262,
      "eval_steps_per_second": 80.128,
      "step": 268
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.5049848556518555,
      "eval_matthews_correlation": 0.43765045412808823,
      "eval_runtime": 0.8256,
      "eval_samples_per_second": 1263.398,
      "eval_steps_per_second": 79.947,
      "step": 402
    },
    {
      "epoch": 3.7313432835820897,
      "grad_norm": 6.807294845581055,
      "learning_rate": 5.515383037871622e-07,
      "loss": 0.4793,
      "step": 500
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.5229844450950623,
      "eval_matthews_correlation": 0.43643553358388004,
      "eval_runtime": 0.9434,
      "eval_samples_per_second": 1105.555,
      "eval_steps_per_second": 69.958,
      "step": 536
    }
  ],
  "logging_steps": 500,
  "max_steps": 536,
  "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": 209747119699272.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 8.211792523053304e-06,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 64,
    "seed": 38
  }
}