{
  "best_metric": 0.5338986272364722,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola\\run-9\\checkpoint-1605",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 2140,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.93,
      "grad_norm": 6.588393688201904,
      "learning_rate": 1.008167296920146e-05,
      "loss": 0.5294,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.4878699779510498,
      "eval_matthews_correlation": 0.47192796522130964,
      "eval_runtime": 0.7915,
      "eval_samples_per_second": 1317.833,
      "eval_steps_per_second": 83.391,
      "step": 535
    },
    {
      "epoch": 1.87,
      "grad_norm": 12.096890449523926,
      "learning_rate": 7.0079921859083314e-06,
      "loss": 0.3753,
      "step": 1000
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.48991820216178894,
      "eval_matthews_correlation": 0.4930285826422816,
      "eval_runtime": 1.0332,
      "eval_samples_per_second": 1009.507,
      "eval_steps_per_second": 63.881,
      "step": 1070
    },
    {
      "epoch": 2.8,
      "grad_norm": 2.1495094299316406,
      "learning_rate": 3.934311402615204e-06,
      "loss": 0.2729,
      "step": 1500
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.5295876264572144,
      "eval_matthews_correlation": 0.5338986272364722,
      "eval_runtime": 0.9835,
      "eval_samples_per_second": 1060.512,
      "eval_steps_per_second": 67.108,
      "step": 1605
    },
    {
      "epoch": 3.74,
      "grad_norm": 25.31566619873047,
      "learning_rate": 8.606306193220758e-07,
      "loss": 0.2181,
      "step": 2000
    },
    {
      "epoch": 4.0,
      "eval_loss": 0.6111654043197632,
      "eval_matthews_correlation": 0.522211073949747,
      "eval_runtime": 0.7971,
      "eval_samples_per_second": 1308.437,
      "eval_steps_per_second": 82.797,
      "step": 2140
    }
  ],
  "logging_steps": 500,
  "max_steps": 2140,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 170994196183080.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 1.3155353752494588e-05,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 16,
    "seed": 15
  }
}