run-4/checkpoint-384/trainer_state.json

{
  "best_metric": 0.5753424657534246,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-384",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 384,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.896482527256012,
      "learning_rate": 2.4429947101690063e-05,
      "loss": 0.3642,
      "step": 192
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.525440313111546,
      "eval_f1": 0.14762741652021089,
      "eval_loss": 0.33992624282836914,
      "eval_mcc": 0.1099564814603399,
      "eval_precision": 0.7241379310344828,
      "eval_recall": 0.0821917808219178,
      "eval_runtime": 66.5832,
      "eval_samples_per_second": 15.349,
      "eval_steps_per_second": 0.481,
      "step": 192
    },
    {
      "epoch": 2.0,
      "grad_norm": 3.451338529586792,
      "learning_rate": 2.1376203713978805e-05,
      "loss": 0.3285,
      "step": 384
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5753424657534246,
      "eval_f1": 0.31761006289308175,
      "eval_loss": 0.31206491589546204,
      "eval_mcc": 0.2299532209711479,
      "eval_precision": 0.808,
      "eval_recall": 0.19765166340508805,
      "eval_runtime": 66.7282,
      "eval_samples_per_second": 15.316,
      "eval_steps_per_second": 0.48,
      "step": 384
    }
  ],
  "logging_steps": 500,
  "max_steps": 1728,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9,
  "save_steps": 500,
  "total_flos": 471390394560.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.4423735218743273,
    "learning_rate": 2.748369048940132e-05,
    "num_train_epochs": 9,
    "per_device_train_batch_size": 16,
    "temperature": 18
  }
}