{
  "best_metric": 0.5422278243727644,
  "best_model_checkpoint": "distilbert-base-uncased-finetuned-cola/run-3/checkpoint-4276",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 8552,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.23,
      "grad_norm": 11.004450798034668,
      "learning_rate": 2.0214687462514598e-05,
      "loss": 0.598,
      "step": 500
    },
    {
      "epoch": 0.47,
      "grad_norm": 15.662084579467773,
      "learning_rate": 1.8959428678205447e-05,
      "loss": 0.5617,
      "step": 1000
    },
    {
      "epoch": 0.7,
      "grad_norm": 28.870677947998047,
      "learning_rate": 1.7704169893896296e-05,
      "loss": 0.5375,
      "step": 1500
    },
    {
      "epoch": 0.94,
      "grad_norm": 64.57504272460938,
      "learning_rate": 1.6448911109587142e-05,
      "loss": 0.5691,
      "step": 2000
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.5517680048942566,
      "eval_matthews_correlation": 0.4384136507504086,
      "eval_runtime": 0.737,
      "eval_samples_per_second": 1415.216,
      "eval_steps_per_second": 89.553,
      "step": 2138
    },
    {
      "epoch": 1.17,
      "grad_norm": 25.73099708557129,
      "learning_rate": 1.5193652325277987e-05,
      "loss": 0.4513,
      "step": 2500
    },
    {
      "epoch": 1.4,
      "grad_norm": 63.40884017944336,
      "learning_rate": 1.3938393540968835e-05,
      "loss": 0.442,
      "step": 3000
    },
    {
      "epoch": 1.64,
      "grad_norm": 0.524051308631897,
      "learning_rate": 1.268313475665968e-05,
      "loss": 0.4237,
      "step": 3500
    },
    {
      "epoch": 1.87,
      "grad_norm": 135.24822998046875,
      "learning_rate": 1.1427875972350528e-05,
      "loss": 0.4295,
      "step": 4000
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.7380757331848145,
      "eval_matthews_correlation": 0.5422278243727644,
      "eval_runtime": 0.7356,
      "eval_samples_per_second": 1417.881,
      "eval_steps_per_second": 89.722,
      "step": 4276
    },
    {
      "epoch": 2.1,
      "grad_norm": 0.07921259105205536,
      "learning_rate": 1.0172617188041375e-05,
      "loss": 0.3849,
      "step": 4500
    },
    {
      "epoch": 2.34,
      "grad_norm": 0.08474227786064148,
      "learning_rate": 8.917358403732224e-06,
      "loss": 0.2999,
      "step": 5000
    },
    {
      "epoch": 2.57,
      "grad_norm": 0.9242565631866455,
      "learning_rate": 7.66209961942307e-06,
      "loss": 0.2708,
      "step": 5500
    },
    {
      "epoch": 2.81,
      "grad_norm": 0.1333150714635849,
      "learning_rate": 6.406840835113916e-06,
      "loss": 0.2846,
      "step": 6000
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.9644601345062256,
      "eval_matthews_correlation": 0.5304969997797532,
      "eval_runtime": 0.7379,
      "eval_samples_per_second": 1413.493,
      "eval_steps_per_second": 89.444,
      "step": 6414
    },
    {
      "epoch": 3.04,
      "grad_norm": 0.13025422394275665,
      "learning_rate": 5.151582050804764e-06,
      "loss": 0.267,
      "step": 6500
    },
    {
      "epoch": 3.27,
      "grad_norm": 0.015170727856457233,
      "learning_rate": 3.896323266495611e-06,
      "loss": 0.1489,
      "step": 7000
    },
    {
      "epoch": 3.51,
      "grad_norm": 0.03684037923812866,
      "learning_rate": 2.641064482186458e-06,
      "loss": 0.1797,
      "step": 7500
    },
    {
      "epoch": 3.74,
      "grad_norm": 0.15966422855854034,
      "learning_rate": 1.385805697877305e-06,
      "loss": 0.1557,
      "step": 8000
    },
    {
      "epoch": 3.98,
      "grad_norm": 0.05443466454744339,
      "learning_rate": 1.3054691356815192e-07,
      "loss": 0.1708,
      "step": 8500
    },
    {
      "epoch": 4.0,
      "eval_loss": 1.0916093587875366,
      "eval_matthews_correlation": 0.5421747077088894,
      "eval_runtime": 0.8194,
      "eval_samples_per_second": 1272.83,
      "eval_steps_per_second": 80.543,
      "step": 8552
    }
  ],
  "logging_steps": 500,
  "max_steps": 8552,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 138657403563900.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 2.1469946246823753e-05,
    "num_train_epochs": 4,
    "per_device_train_batch_size": 4,
    "seed": 10
  }
}