config.json

{
  "model_type": "attention_unet",
  "backbone": "convnext",
  "img_size": 256,
  "in_channels": 3,
  "out_channels": 1,
  "auto_map": {
    "AutoConfig": "modeling_attention_unet.AttentionUNetConfig",
    "AutoModel": "modeling_attention_unet.AttentionUNetForSegmentation"
  },
  "training": {
    "loss": "bce_dice",
    "lr": 0.001,
    "weight_decay": 0.0001,
    "warmup_ratio": 0.05,
    "scheduler": "cosine",
    "batch_size": 64,
    "epochs": 50,
    "optimiser": "AdamW"
  },
  "test_metrics": {
    "dice": 0.9411258697509766,
    "iou": 0.8887985348701477,
    "f1": 0.94112581440282,
    "precision": 0.9617935460534044,
    "recall": 0.9213276440087947,
    "accuracy": 0.9803045654296875,
    "loss": 0.0876474380493164
  },
  "dataset": "andreribeiro87/kvasir-seg-augmented",
  "note": "Best test-Dice model from architecture sweep (no HPO applied)"
}