WoundAmbit: Bridging State-of-the-Art Semantic Segmentation and Real-World Wound Care (Trained K-Fold Models)

The provided zip-folder contains all trained k-fold models used for our publication.
Each of the 12 models (fcbformer, fusegnet, hardnetdfus, hiformerb, internimageupernet_t, missformer, segformerb3, segnextl, transnextupernet_tiny, unet, vwformerconvnexts, vwformermitb3) follows the same structure:

📂 <model_name>
 ├── 📁 fold1
 │   └── best_checkpoint.pt
 ├── 📁 fold2
 │   └── best_checkpoint.pt
 ├── 📁 fold3
 │   └── best_checkpoint.pt
 ├── 📁 fold4
 │   └── best_checkpoint.pt
 ├── 📁 fold5
 │   └── best_checkpoint.pt
 ├── <model_name>-CFU-512-k-Fold.yaml
 └── k_fold_state.pkl