Semantic Segmentation on ADE20k 92 (val)
62.8mIoUAbs. SOTA
Evaluation results
| Method | Links | |
|---|---|---|
| Abs. SOTAReference=SOTA from [77]2025.03 | 62.8 | |
| Original M2F2025.03 | 57.7 | |
| DINOv3 + DA(SDP)Adapter=SDP2025.03 | 54.7 | |
| DINOv2 + DA(No Perlin)Adapter=No Perlin2025.03 | 54 | |
| DINOv2 + DA(Norm)Adapter=Norm2025.03 | 53.6 | |
| DINOv2 + DA(SDP)Adapter=SDP2025.03 | 53.4 | |
| DINOv22025.03 | 51.7 | |
| DepthAnythingV2 + DinoV22025.03 | 49.7 | |
| DINOv32025.03 | 48.9 | |
| DepthAnythingV22025.03 | 48.7 | |
| Omnivore2025.03 | 45.9 | |
| DFormerV1-pre2025.03 | 44.7 | |
| Omnivore + Depth2025.03 | 44.3 | |
| EVA-02 + DA(SDP)Adapter=SDP2025.03 | 43.5 | |
| Dune2025.03 | 42.9 | |
| DFormerV2-pre2025.03 | 41.4 | |
| MultiMAE + Depth2025.03 | 41.1 | |
| FiT3D2025.03 | 40.5 | |
| OmniDC-pre2025.03 | 36.9 | |
| DA(No pre-training) + DA(SDP)Adapter=SDP2025.03 | 36.6 | |
| DMD3C-pre2025.03 | 36.3 | |
| MASt3R (ViT-L)2025.03 | 33.1 | |
| EVA-022025.03 | 32.7 | |
| MultiMAE2025.03 | 30.8 | |
| CroCo V22025.03 | 30.2 | |
| DA(No pre-training)2025.03 | 29.5 |