run-42/checkpoint-1581/trainer_state.json

{
  "best_metric": 0.8405963302752294,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-42/checkpoint-527",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 1581,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 12.218423843383789,
      "learning_rate": 0.0002468577923854863,
      "loss": 1.1492,
      "step": 527
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.8405963302752294,
      "eval_loss": 1.2373847961425781,
      "eval_runtime": 2.8121,
      "eval_samples_per_second": 310.092,
      "eval_steps_per_second": 2.489,
      "step": 527
    },
    {
      "epoch": 2.0,
      "grad_norm": 7.051322937011719,
      "learning_rate": 0.00012342889619274314,
      "loss": 0.5283,
      "step": 1054
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8302752293577982,
      "eval_loss": 1.383708119392395,
      "eval_runtime": 2.8134,
      "eval_samples_per_second": 309.949,
      "eval_steps_per_second": 2.488,
      "step": 1054
    },
    {
      "epoch": 3.0,
      "grad_norm": NaN,
      "learning_rate": 2.3421042920824125e-07,
      "loss": 0.3751,
      "step": 1581
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.8188073394495413,
      "eval_loss": 1.4459969997406006,
      "eval_runtime": 2.8106,
      "eval_samples_per_second": 310.249,
      "eval_steps_per_second": 2.491,
      "step": 1581
    }
  ],
  "logging_steps": 500,
  "max_steps": 1581,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 24228458327580.0,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.47037947579293116,
    "learning_rate": 0.00037028668857822943,
    "num_train_epochs": 3,
    "temperature": 12
  }
}