run-48/checkpoint-384/trainer_state.json

{
  "best_metric": 0.8101761252446184,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-48/checkpoint-384",
  "epoch": 2.0,
  "eval_steps": 500,
  "global_step": 384,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 6.156297206878662,
      "learning_rate": 6.847131652416489e-05,
      "loss": 0.5617,
      "step": 192
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7984344422700587,
      "eval_f1": 0.8186619718309858,
      "eval_loss": 0.4545396566390991,
      "eval_mcc": 0.6123004481387646,
      "eval_precision": 0.744,
      "eval_recall": 0.9099804305283757,
      "eval_runtime": 66.7324,
      "eval_samples_per_second": 15.315,
      "eval_steps_per_second": 0.48,
      "step": 192
    },
    {
      "epoch": 2.0,
      "grad_norm": 6.695183277130127,
      "learning_rate": 0.0,
      "loss": 0.4646,
      "step": 384
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8101761252446184,
      "eval_f1": 0.8245931283905967,
      "eval_loss": 0.43737441301345825,
      "eval_mcc": 0.6289075983671709,
      "eval_precision": 0.7663865546218488,
      "eval_recall": 0.8923679060665362,
      "eval_runtime": 66.9568,
      "eval_samples_per_second": 15.264,
      "eval_steps_per_second": 0.478,
      "step": 384
    }
  ],
  "logging_steps": 500,
  "max_steps": 384,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 2,
  "save_steps": 500,
  "total_flos": 471390394560.0,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": {
    "alpha": 0.8915947606865525,
    "learning_rate": 0.00013694263304832977,
    "num_train_epochs": 2,
    "per_device_train_batch_size": 16,
    "temperature": 24
  }
}