run-11/checkpoint-384/trainer_state.json

{
  "best_metric": 0.7857142857142857,
  "best_model_checkpoint": "tiny-bert-sst2-distilled/run-11/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": 3.698505163192749,
      "learning_rate": 0.0004729463641993846,
      "loss": 0.4978,
      "step": 192
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7465753424657534,
      "eval_f1": 0.7821698906644239,
      "eval_loss": 0.46408411860466003,
      "eval_mcc": 0.5218028809356926,
      "eval_precision": 0.6858407079646017,
      "eval_recall": 0.9099804305283757,
      "eval_runtime": 66.6598,
      "eval_samples_per_second": 15.332,
      "eval_steps_per_second": 0.48,
      "step": 192
    },
    {
      "epoch": 2.0,
      "grad_norm": 4.018568515777588,
      "learning_rate": 0.0,
      "loss": 0.4526,
      "step": 384
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7857142857142857,
      "eval_f1": 0.7999999999999999,
      "eval_loss": 0.43016862869262695,
      "eval_mcc": 0.5773502691896257,
      "eval_precision": 0.75,
      "eval_recall": 0.8571428571428571,
      "eval_runtime": 66.4246,
      "eval_samples_per_second": 15.386,
      "eval_steps_per_second": 0.482,
      "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": 0.7646810195593648,
    "learning_rate": 0.0009458927283987692,
    "num_train_epochs": 2,
    "per_device_train_batch_size": 16,
    "temperature": 27
  }
}