{
  "optimal_threshold": 0.5,
  "test_results": {
    "eval_loss": 0.16758447885513306,
    "eval_accuracy": 0.9857604679802956,
    "eval_f1": 0.9332852506310855,
    "eval_precision": 0.9635145197319435,
    "eval_recall": 0.9048951048951049,
    "eval_roc_auc": 0.9925474758764847,
    "eval_class_metrics": {
      "class_0_precision": 0.9883251781268778,
      "class_0_recall": 0.9957619788963847,
      "class_0_f1": 0.9920296411184353,
      "class_1_precision": 0.9635145197319435,
      "class_1_recall": 0.9048951048951049,
      "class_1_f1": 0.9332852506310855
    },
    "eval_runtime": 220.015,
    "eval_samples_per_second": 59.051,
    "eval_steps_per_second": 1.845,
    "epoch": 0.22621591051903983
  },
  "eval_results": {
    "eval_loss": 0.12878268957138062,
    "eval_accuracy": 0.9872229064039408,
    "eval_f1": 0.9404591104734576,
    "eval_precision": 0.9632623071271125,
    "eval_recall": 0.9187105816398038,
    "eval_roc_auc": 0.9958163404734399,
    "eval_class_metrics": {
      "class_0_precision": 0.9900266529103259,
      "class_0_recall": 0.9956766104626027,
      "class_0_f1": 0.9928435937230558,
      "class_1_precision": 0.9632623071271125,
      "class_1_recall": 0.9187105816398038,
      "class_1_f1": 0.9404591104734576
    },
    "eval_runtime": 222.6834,
    "eval_samples_per_second": 58.343,
    "eval_steps_per_second": 1.823,
    "epoch": 0.22621591051903983
  },
  "loss_function": "cross_entropy",
  "loss_params": {
    "margin_in": 2.0,
    "margin_out": 18.0,
    "temperature": 1.0
  },
  "params": {
    "base_model": "jhu-clsp/mmBERT-base",
    "model_id": "modernguard-mmBERT-base-4-cross_entropy",
    "huggingface_repo": null,
    "dataset_id": "rafaelsandroni/modern-guard-v4-training",
    "max_length": 2048,
    "stride": 128,
    "batch_size": 8,
    "epochs": 20,
    "learning_rate": 1e-05,
    "weight_decay": 0.05,
    "warmup_ratio": 0.1,
    "margin_in": 2.0,
    "margin_out": 18.0,
    "eval_steps": 500,
    "early_stopping_patience": 3,
    "early_stopping_threshold": 0.001,
    "loss_function": "cross_entropy",
    "temperature": 1.0,
    "energy_weight": 0.5,
    "focal_gamma": 2.0,
    "use_class_weights": true,
    "seed": 42,
    "max_samples": -1
  },
  "tokenization_type": "stride-based",
  "dataset_stats": {
    "original_examples": 1296302,
    "train_original": 1270375,
    "val_original": 12963,
    "test_original": 12964
  }
}