run-4/checkpoint-594/trainer_state.json

{
  "best_metric": 0.594059405940594,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-594",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 594,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.202705979347229,
      "learning_rate": 6.504485936113838e-05,
      "loss": 0.5495,
      "step": 297
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5069306930693069,
      "eval_f1": 0.10108303249097472,
      "eval_loss": 0.5468167662620544,
      "eval_mcc": 0.027838098756040194,
      "eval_precision": 0.56,
      "eval_recall": 0.05555555555555555,
      "eval_runtime": 0.9316,
      "eval_samples_per_second": 542.063,
      "eval_steps_per_second": 17.174,
      "step": 297
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8953370451927185,
      "learning_rate": 5.4204049467615325e-05,
      "loss": 0.54,
      "step": 594
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.594059405940594,
      "eval_f1": 0.5858585858585857,
      "eval_loss": 0.5394836068153381,
      "eval_mcc": 0.18817791261380143,
      "eval_precision": 0.5967078189300411,
      "eval_recall": 0.5753968253968254,
      "eval_runtime": 0.9335,
      "eval_samples_per_second": 540.967,
      "eval_steps_per_second": 17.14,
      "step": 594
    }
  ],
  "logging_steps": 500,
  "max_steps": 2079,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 1461402531960.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.7404813991868276,
    "learning_rate": 7.588566925466145e-05,
    "num_train_epochs": 7,
    "temperature": 47
  }
}