run-2/checkpoint-144/trainer_state.json

{
  "best_metric": 0.5,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-2/checkpoint-48",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 144,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.6336328983306885,
      "learning_rate": 8.611669739559475e-06,
      "loss": 0.3627,
      "step": 48
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5,
      "eval_f1": 0.0,
      "eval_loss": 0.31828778982162476,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 28.8885,
      "eval_samples_per_second": 35.377,
      "eval_steps_per_second": 0.554,
      "step": 48
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.23749461770057678,
      "learning_rate": 4.3058348697797375e-06,
      "loss": 0.3092,
      "step": 96
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.5,
      "eval_f1": 0.0,
      "eval_loss": 0.29845258593559265,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 28.7105,
      "eval_samples_per_second": 35.597,
      "eval_steps_per_second": 0.557,
      "step": 96
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.9887090921401978,
      "learning_rate": 0.0,
      "loss": 0.2993,
      "step": 144
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.5,
      "eval_f1": 0.0,
      "eval_loss": 0.29658088088035583,
      "eval_precision": 0.0,
      "eval_recall": 0.0,
      "eval_runtime": 29.7133,
      "eval_samples_per_second": 34.395,
      "eval_steps_per_second": 0.538,
      "step": 144
    }
  ],
  "logging_steps": 500,
  "max_steps": 144,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 707085591840.0,
  "train_batch_size": 64,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.340965759348954,
    "learning_rate": 1.2917504609339214e-05,
    "num_train_epochs": 3,
    "temperature": 24
  }
}