{
  "best_metric": 0.8371559633027523,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-34/checkpoint-527",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 3689,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 20.85869789123535,
      "learning_rate": 0.0002713055538909492,
      "loss": 1.5596,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8371559633027523,
      "eval_loss": 1.5567338466644287,
      "eval_runtime": 2.8111,
      "eval_samples_per_second": 310.197,
      "eval_steps_per_second": 2.49,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 6.722352027893066,
      "learning_rate": 0.00022608796157579106,
      "loss": 0.6915,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.819954128440367,
      "eval_loss": 1.7321298122406006,
      "eval_runtime": 2.823,
      "eval_samples_per_second": 308.894,
      "eval_steps_per_second": 2.48,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": NaN,
      "learning_rate": 0.0001809561711435838,
      "loss": 0.4888,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.819954128440367,
      "eval_loss": 1.7588660717010498,
      "eval_runtime": 2.8219,
      "eval_samples_per_second": 309.007,
      "eval_steps_per_second": 2.481,
      "step": 1581
    },
    {
      "epoch": 4.0,
      "grad_norm": 5.12546443939209,
      "learning_rate": 0.0001357385788284256,
      "loss": 0.3746,
      "step": 2108
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8279816513761468,
      "eval_loss": 1.7675563097000122,
      "eval_runtime": 2.821,
      "eval_samples_per_second": 309.105,
      "eval_steps_per_second": 2.481,
      "step": 2108
    },
    {
      "epoch": 5.0,
      "grad_norm": 11.028658866882324,
      "learning_rate": 9.052098651326739e-05,
      "loss": 0.2954,
      "step": 2635
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8211009174311926,
      "eval_loss": 1.831598162651062,
      "eval_runtime": 2.8049,
      "eval_samples_per_second": 310.88,
      "eval_steps_per_second": 2.496,
      "step": 2635
    },
    {
      "epoch": 6.0,
      "grad_norm": 19.683809280395508,
      "learning_rate": 4.5303394198109176e-05,
      "loss": 0.2513,
      "step": 3162
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.8222477064220184,
      "eval_loss": 1.8130003213882446,
      "eval_runtime": 2.8237,
      "eval_samples_per_second": 308.816,
      "eval_steps_per_second": 2.479,
      "step": 3162
    },
    {
      "epoch": 7.0,
      "grad_norm": 4.766151428222656,
      "learning_rate": 8.580188295096434e-08,
      "loss": 0.2259,
      "step": 3689
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.8165137614678899,
      "eval_loss": 1.8347537517547607,
      "eval_runtime": 2.8047,
      "eval_samples_per_second": 310.903,
      "eval_steps_per_second": 2.496,
      "step": 3689
    }
  ],
  "logging_steps": 500,
  "max_steps": 3689,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 56628612372420.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.2366804151203985,
    "learning_rate": 0.0003165231462061075,
    "num_train_epochs": 7,
    "temperature": 22
  }
}