run-38/checkpoint-856/trainer_state.json

{
  "best_metric": 0.7521968365553603,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-38/checkpoint-642",
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 856,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 5.823625087738037,
      "learning_rate": 0.00022026694526910956,
      "loss": 0.4997,
      "step": 214
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7129466900995899,
      "eval_f1": 0.40243902439024387,
      "eval_loss": 0.4626123011112213,
      "eval_mcc": 0.2854021761857005,
      "eval_precision": 0.6573705179282868,
      "eval_recall": 0.28998242530755713,
      "eval_runtime": 3.1456,
      "eval_samples_per_second": 542.661,
      "eval_steps_per_second": 17.167,
      "step": 214
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.7993404865264893,
      "learning_rate": 0.0001888002388020939,
      "loss": 0.4542,
      "step": 428
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7398945518453427,
      "eval_f1": 0.47641509433962265,
      "eval_loss": 0.4441252648830414,
      "eval_mcc": 0.366324992569054,
      "eval_precision": 0.7240143369175627,
      "eval_recall": 0.35500878734622143,
      "eval_runtime": 3.7126,
      "eval_samples_per_second": 459.782,
      "eval_steps_per_second": 14.545,
      "step": 428
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.832775354385376,
      "learning_rate": 0.00015733353233507826,
      "loss": 0.4371,
      "step": 642
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7521968365553603,
      "eval_f1": 0.5052631578947367,
      "eval_loss": 0.44420263171195984,
      "eval_mcc": 0.4015258565021214,
      "eval_precision": 0.7552447552447552,
      "eval_recall": 0.37961335676625657,
      "eval_runtime": 3.1594,
      "eval_samples_per_second": 540.296,
      "eval_steps_per_second": 17.092,
      "step": 642
    },
    {
      "epoch": 4.0,
      "grad_norm": 2.8890116214752197,
      "learning_rate": 0.0001258668258680626,
      "loss": 0.4321,
      "step": 856
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7398945518453427,
      "eval_f1": 0.43367346938775514,
      "eval_loss": 0.4393432140350342,
      "eval_mcc": 0.3683014058572899,
      "eval_precision": 0.7906976744186046,
      "eval_recall": 0.29876977152899825,
      "eval_runtime": 3.8868,
      "eval_samples_per_second": 439.178,
      "eval_steps_per_second": 13.893,
      "step": 856
    }
  ],
  "logging_steps": 500,
  "max_steps": 1712,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 2099102657760.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.806613569484766,
    "learning_rate": 0.0002517336517361252,
    "num_train_epochs": 8,
    "temperature": 7
  }
}