run-8/checkpoint-66/trainer_state.json

{
  "best_metric": 0.5078277886497065,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-8/checkpoint-66",
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 66,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 2.3361287117004395,
      "learning_rate": 1.2926171014641113e-05,
      "loss": 0.6798,
      "step": 66
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5078277886497065,
      "eval_f1": 0.16583747927031509,
      "eval_loss": 0.6719066500663757,
      "eval_precision": 0.5434782608695652,
      "eval_recall": 0.09784735812133072,
      "eval_runtime": 27.922,
      "eval_samples_per_second": 36.602,
      "eval_steps_per_second": 1.146,
      "step": 66
    }
  ],
  "logging_steps": 500,
  "max_steps": 462,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 500,
  "total_flos": 235695197280.0,
  "train_batch_size": 47,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.9637337898381267,
    "learning_rate": 1.5080532850414631e-05,
    "num_train_epochs": 7,
    "per_device_train_batch_size": 47,
    "temperature": 3
  }
}