run-16/checkpoint-320/trainer_state.json

{
  "best_metric": 0.7137254901960784,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-16/checkpoint-320",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 320,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 1.7216335535049438,
      "learning_rate": 0.00033371467549721734,
      "loss": 0.4146,
      "step": 160
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6870588235294117,
      "eval_f1": 0.1971830985915493,
      "eval_loss": 0.37955978512763977,
      "eval_mcc": 0.18019685365777874,
      "eval_precision": 0.6805555555555556,
      "eval_recall": 0.11529411764705882,
      "eval_runtime": 1.871,
      "eval_samples_per_second": 681.446,
      "eval_steps_per_second": 21.379,
      "step": 160
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.9292495250701904,
      "learning_rate": 0.00028604115042618627,
      "loss": 0.3901,
      "step": 320
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7137254901960784,
      "eval_f1": 0.33515482695810567,
      "eval_loss": 0.3745056092739105,
      "eval_mcc": 0.2844985686324911,
      "eval_precision": 0.7419354838709677,
      "eval_recall": 0.2164705882352941,
      "eval_runtime": 1.8722,
      "eval_samples_per_second": 681.009,
      "eval_steps_per_second": 21.365,
      "step": 320
    }
  ],
  "logging_steps": 500,
  "max_steps": 1280,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 8,
  "save_steps": 500,
  "total_flos": 583510875840.0,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.640212771663602,
    "learning_rate": 0.0003813882005682484,
    "num_train_epochs": 8,
    "temperature": 30
  }
}