run-35/checkpoint-297/trainer_state.json

{
  "best_metric": 0.497029702970297,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-35/checkpoint-297",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 297,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.011061429977417,
      "learning_rate": 0.0008596139358612845,
      "loss": 0.6951,
      "step": 297
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.497029702970297,
      "eval_f1": 0.664021164021164,
      "eval_loss": 0.6919692158699036,
      "eval_mcc": -0.044631832853976164,
      "eval_precision": 0.498015873015873,
      "eval_recall": 0.996031746031746,
      "eval_runtime": 0.9311,
      "eval_samples_per_second": 542.385,
      "eval_steps_per_second": 17.184,
      "step": 297
    }
  ],
  "logging_steps": 500,
  "max_steps": 2376,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 730701265980.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9987048243469214,
    "learning_rate": 0.0009824159266986108,
    "num_train_epochs": 8,
    "temperature": 5
  }
}