| Benchmark | Model | Rank | Results |
|---|---|---|---|
| object-detection-on-coco-val2017 | Hyb-KAN ViT-B (Mask R-CNN, ViTDet-style, 3x) | – | box AP: 50.6AP50: 71.3AP75: 55.3Params (M): 140 |
| object-detection-on-coco-val2017 | Hyb-KAN ViT-S (Mask R-CNN, ViTDet-style, 3x) | – | box AP: 48.3AP50: 69.8AP75: 52.7Params (M): 51.7 |