{
  "best_metric": 0.72,
  "best_model_checkpoint": "distilhubert-finetuned-not-a-word2/run-4/checkpoint-48",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 48,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "grad_norm": 1.2733114957809448,
      "learning_rate": 2.2702186710865246e-07,
      "loss": 0.7025,
      "step": 5
    },
    {
      "epoch": 0.21,
      "grad_norm": 1.243804931640625,
      "learning_rate": 4.5404373421730493e-07,
      "loss": 0.6974,
      "step": 10
    },
    {
      "epoch": 0.31,
      "grad_norm": 1.7711552381515503,
      "learning_rate": 6.810656013259573e-07,
      "loss": 0.696,
      "step": 15
    },
    {
      "epoch": 0.42,
      "grad_norm": 1.1453403234481812,
      "learning_rate": 9.080874684346099e-07,
      "loss": 0.6989,
      "step": 20
    },
    {
      "epoch": 0.52,
      "grad_norm": 1.2729355096817017,
      "learning_rate": 1.1351093355432624e-06,
      "loss": 0.6968,
      "step": 25
    },
    {
      "epoch": 0.62,
      "grad_norm": 1.1592165231704712,
      "learning_rate": 1.3621312026519146e-06,
      "loss": 0.6959,
      "step": 30
    },
    {
      "epoch": 0.73,
      "grad_norm": 1.1798148155212402,
      "learning_rate": 1.589153069760567e-06,
      "loss": 0.6952,
      "step": 35
    },
    {
      "epoch": 0.83,
      "grad_norm": 2.1216671466827393,
      "learning_rate": 1.8161749368692197e-06,
      "loss": 0.6886,
      "step": 40
    },
    {
      "epoch": 0.94,
      "grad_norm": 1.3416370153427124,
      "learning_rate": 2.043196803977872e-06,
      "loss": 0.6864,
      "step": 45
    },
    {
      "epoch": 1.0,
      "eval_f1": 0.72,
      "eval_loss": 0.688262939453125,
      "eval_runtime": 1.3468,
      "eval_samples_per_second": 47.521,
      "eval_steps_per_second": 5.94,
      "step": 48
    }
  ],
  "logging_steps": 5,
  "max_steps": 480,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 670686130935120.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 2.1794099242430636e-06,
    "per_device_train_batch_size": 4
  }
}