run-20/checkpoint-96/trainer_state.json

{
  "best_metric": 0.8170254403131115,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-20/checkpoint-96",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 96,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.930127739906311,
      "learning_rate": 0.00038924925342110863,
      "loss": 0.5828,
      "step": 48
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7553816046966731,
      "eval_f1": 0.7967479674796748,
      "eval_loss": 0.4942302703857422,
      "eval_precision": 0.6815020862308763,
      "eval_recall": 0.958904109589041,
      "eval_runtime": 27.5665,
      "eval_samples_per_second": 37.074,
      "eval_steps_per_second": 0.58,
      "step": 48
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.047781944274902,
      "learning_rate": 0.0002594995022807391,
      "loss": 0.4792,
      "step": 96
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8170254403131115,
      "eval_f1": 0.8270120259019426,
      "eval_loss": 0.4411511719226837,
      "eval_precision": 0.7842105263157895,
      "eval_recall": 0.8747553816046967,
      "eval_runtime": 27.2029,
      "eval_samples_per_second": 37.57,
      "eval_steps_per_second": 0.588,
      "step": 96
    }
  ],
  "logging_steps": 500,
  "max_steps": 192,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 500,
  "total_flos": 471390394560.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9873470402997029,
    "learning_rate": 0.0005189990045614782,
    "num_train_epochs": 4,
    "temperature": 29
  }
}