{
  "train_metrics": {
    "train_runtime": 397.6795,
    "train_samples_per_second": 37.316,
    "train_steps_per_second": 1.169,
    "total_flos": 3123710468358144.0,
    "train_loss": 0.2441582655714404,
    "epoch": 4.0
  },
  "eval_metrics": {
    "eval_loss": 0.08734048902988434,
    "eval_accuracy": 0.9669291338582677,
    "eval_precision": 0.9759615384615384,
    "eval_recall": 0.973621103117506,
    "eval_f1": 0.9747899159663865,
    "eval_auroc": 0.9965018810639561,
    "eval_fpr": 0.045871559633027525,
    "eval_runtime": 6.9183,
    "eval_samples_per_second": 91.786,
    "eval_steps_per_second": 11.564,
    "epoch": 4.0
  },
  "config": {
    "model": {
      "name": "microsoft/deberta-v3-base",
      "max_seq_length": 512
    },
    "training": {
      "num_train_epochs": 5,
      "per_device_train_batch_size": 16,
      "gradient_accumulation_steps": 2,
      "learning_rate": 2e-05,
      "lr_scheduler_type": "cosine",
      "warmup_ratio": 0.1,
      "weight_decay": 0.01,
      "bf16": false,
      "seed": 42,
      "logging_steps": 50,
      "eval_strategy": "epoch",
      "save_strategy": "epoch",
      "save_total_limit": 3,
      "load_best_model_at_end": true,
      "metric_for_best_model": "f1",
      "early_stopping_patience": 2,
      "optim": "adamw_torch"
    },
    "data": {
      "train_file": "data/processed/train.jsonl",
      "val_file": "data/processed/val.jsonl",
      "test_file": "data/processed/test.jsonl",
      "class_weights": true,
      "text_column": "text",
      "label_column": "label"
    },
    "output": {
      "model_dir": "models/deberta_bioguard_v1",
      "log_dir": "results/training_logs"
    }
  }
}