run-12/checkpoint-384/trainer_state.json

{
  "best_metric": 0.7954990215264188,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-12/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": 0.9969735741615295,
      "learning_rate": 0.0003519988357465585,
      "loss": 0.4838,
      "step": 192
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7915851272015656,
      "eval_f1": 0.8149435273675064,
      "eval_loss": 0.4351368844509125,
      "eval_mcc": 0.6026908465743466,
      "eval_precision": 0.7328125,
      "eval_recall": 0.9178082191780822,
      "eval_runtime": 66.7179,
      "eval_samples_per_second": 15.318,
      "eval_steps_per_second": 0.48,
      "step": 192
    },
    {
      "epoch": 2.0,
      "grad_norm": 9.071135520935059,
      "learning_rate": 0.0002933323631221321,
      "loss": 0.435,
      "step": 384
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7954990215264188,
      "eval_f1": 0.7956989247311829,
      "eval_loss": 0.42000794410705566,
      "eval_mcc": 0.5909991747116098,
      "eval_precision": 0.794921875,
      "eval_recall": 0.7964774951076321,
      "eval_runtime": 66.4436,
      "eval_samples_per_second": 15.381,
      "eval_steps_per_second": 0.482,
      "step": 384
    }
  ],
  "logging_steps": 500,
  "max_steps": 1344,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 471390394560.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.7654129557516562,
    "learning_rate": 0.00041066530837098494,
    "num_train_epochs": 7,
    "per_device_train_batch_size": 16,
    "temperature": 40
  }
}