{
  "best_metric": 0.7632093933463796,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-2/checkpoint-861",
  "epoch": 7.0,
  "eval_steps": 500,
  "global_step": 861,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.0383042097091675,
      "learning_rate": 2.0778074552979435e-05,
      "loss": 0.4799,
      "step": 123
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5,
      "eval_f1": 0.0,
      "eval_loss": 0.4709824323654175,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 29.6795,
      "eval_samples_per_second": 34.435,
      "eval_steps_per_second": 1.078,
      "step": 123
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.8932008743286133,
      "learning_rate": 1.780977818826809e-05,
      "loss": 0.4618,
      "step": 246
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5322896281800391,
      "eval_f1": 0.1758620689655172,
      "eval_loss": 0.44377976655960083,
      "eval_precision": 0.7391304347826086,
      "eval_recall": 0.09980430528375733,
      "eval_runtime": 29.7432,
      "eval_samples_per_second": 34.361,
      "eval_steps_per_second": 1.076,
      "step": 246
    },
    {
      "epoch": 3.0,
      "grad_norm": 2.3748350143432617,
      "learning_rate": 1.484148182355674e-05,
      "loss": 0.4269,
      "step": 369
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.6545988258317026,
      "eval_f1": 0.5456885456885457,
      "eval_loss": 0.4057740271091461,
      "eval_precision": 0.7969924812030075,
      "eval_recall": 0.41487279843444225,
      "eval_runtime": 29.4826,
      "eval_samples_per_second": 34.665,
      "eval_steps_per_second": 1.085,
      "step": 369
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.1657488346099854,
      "learning_rate": 1.1873185458845392e-05,
      "loss": 0.407,
      "step": 492
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7113502935420744,
      "eval_f1": 0.671857619577308,
      "eval_loss": 0.39306214451789856,
      "eval_precision": 0.7783505154639175,
      "eval_recall": 0.5909980430528375,
      "eval_runtime": 29.5736,
      "eval_samples_per_second": 34.558,
      "eval_steps_per_second": 1.082,
      "step": 492
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.3129783868789673,
      "learning_rate": 8.904889094134044e-06,
      "loss": 0.4011,
      "step": 615
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7270058708414873,
      "eval_f1": 0.7053854276663145,
      "eval_loss": 0.38719648122787476,
      "eval_precision": 0.7660550458715596,
      "eval_recall": 0.6536203522504892,
      "eval_runtime": 30.8479,
      "eval_samples_per_second": 33.13,
      "eval_steps_per_second": 1.037,
      "step": 615
    },
    {
      "epoch": 6.0,
      "grad_norm": 2.9716053009033203,
      "learning_rate": 5.936592729422696e-06,
      "loss": 0.3941,
      "step": 738
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.7553816046966731,
      "eval_f1": 0.7469635627530365,
      "eval_loss": 0.38426536321640015,
      "eval_precision": 0.7735849056603774,
      "eval_recall": 0.7221135029354208,
      "eval_runtime": 29.4509,
      "eval_samples_per_second": 34.702,
      "eval_steps_per_second": 1.087,
      "step": 738
    },
    {
      "epoch": 7.0,
      "grad_norm": 3.693033218383789,
      "learning_rate": 2.968296364711348e-06,
      "loss": 0.3958,
      "step": 861
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.7632093933463796,
      "eval_f1": 0.7584830339321357,
      "eval_loss": 0.38294512033462524,
      "eval_precision": 0.7739307535641547,
      "eval_recall": 0.7436399217221135,
      "eval_runtime": 29.3807,
      "eval_samples_per_second": 34.785,
      "eval_steps_per_second": 1.089,
      "step": 861
    }
  ],
  "logging_steps": 500,
  "max_steps": 984,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 1649866380960.0,
  "train_batch_size": 25,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.6164141606295046,
    "learning_rate": 2.3746370917690784e-05,
    "num_train_epochs": 8,
    "per_device_train_batch_size": 25,
    "temperature": 15
  }
}