run-10/checkpoint-128/trainer_state.json

{
  "best_metric": 0.799412915851272,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-10/checkpoint-128",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 128,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 7.969868183135986,
      "learning_rate": 0.0006324213595544264,
      "loss": 0.5787,
      "step": 64
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.773972602739726,
      "eval_f1": 0.8070175438596492,
      "eval_loss": 0.48913490772247314,
      "eval_precision": 0.7040816326530612,
      "eval_recall": 0.9452054794520548,
      "eval_runtime": 27.7283,
      "eval_samples_per_second": 36.858,
      "eval_steps_per_second": 1.154,
      "step": 64
    },
    {
      "epoch": 2.0,
      "grad_norm": 2.9250471591949463,
      "learning_rate": 0.00047431601966581977,
      "loss": 0.46,
      "step": 128
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.799412915851272,
      "eval_f1": 0.8138056312443234,
      "eval_loss": 0.4400319457054138,
      "eval_precision": 0.7593220338983051,
      "eval_recall": 0.8767123287671232,
      "eval_runtime": 28.2626,
      "eval_samples_per_second": 36.161,
      "eval_steps_per_second": 1.132,
      "step": 128
    }
  ],
  "logging_steps": 500,
  "max_steps": 320,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 471390394560.0,
  "train_batch_size": 48,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9990297556295253,
    "learning_rate": 0.000790526699443033,
    "num_train_epochs": 5,
    "per_device_train_batch_size": 48,
    "temperature": 30
  }
}