Semantic Segmentation on Dv 58-class (val)
62ACDC-41S2-Corr
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| S2-CorrBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 62 | 61.7 | 47.4 | 41.9 | 53.2 | |
| S2-CorrBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 62 | 61.7 | 47.4 | 41.9 | 53.2 | |
| CLIPSelfBackbone=EVA02 ViT-L/14, Training Data=CS-7+COCO2026.02 | 61.3 | 60.3 | 44.7 | 39.8 | 51.5 | |
| CLIPSelfBackbone=EVA02 ViT-L/14, Training Data=CS-7+CoCo2026.02 | 61.3 | 60.3 | 44.7 | 39.8 | 51.5 | |
| RSC-CLIPSelfBackbone=EVA02 ViT-L/14, Training Data=CS-7+CoCo2026.02 | 61.3 | 59.6 | 44.6 | 36.4 | 50.5 | |
| CAT-Seg+DGInStyleBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 60.2 | 60.8 | 43.5 | 38.7 | 50.8 | |
| CAT-Seg+DGInStyleBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 60.2 | 60.8 | 43.5 | 38.7 | 50.8 | |
| MaskAdapterBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 59.7 | 59.7 | 41.8 | 36.1 | 49.3 | |
| MaskAdapterBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 59.7 | 59.7 | 41.8 | 36.1 | 49.3 | |
| CAT-SegBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 59.2 | 60 | 41.1 | 39.5 | 50 | |
| CAT-SegBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 59.2 | 60 | 41.1 | 39.5 | 50 | |
| MAFT+Backbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 58.4 | 57.9 | 42.2 | 39 | 49.4 | |
| SANBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 57.5 | 56.9 | 38 | 34.8 | 46.8 | |
| ESC-NetBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 57.1 | 59.2 | 43.8 | 36.5 | 49.2 | |
| CAT-Seg+AdvStyleBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 56 | 56.4 | 40.8 | 38.6 | 48 | |
| S2-CorrBackbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 55.4 | 59.8 | 41.5 | 40.8 | 49.4 | |
| CAT-Seg+DGInStyleBackbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 53.7 | 59 | 37.9 | 38 | 47.2 | |
| CAT-SegBackbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 53.6 | 59.7 | 41.1 | 38.2 | 48.2 | |
| S2-CorrBackbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 52.3 | 58.6 | 42 | 38.6 | 47.9 | |
| MaskAdapterBackbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 52.1 | 55.9 | 37.1 | 37.1 | 45.6 | |
| RSC-CLIPSelfBackbone=EVA02 ViT-L/14, Training Data=GTA-7+CoCo2026.02 | 52 | 59.1 | 39.8 | 38.4 | 47.8 | |
| CAT-Seg+AdvStyleBackbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 51.8 | 59.2 | 40.7 | 35.5 | 46.8 | |
| ESC-NetBackbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 51.5 | 57.3 | 38.7 | 35.1 | 45.7 | |
| OVSegBackbone=EVA02 ViT-L/14, Training Data=CS-72026.02 | 51.2 | 49.3 | 33.7 | 35.7 | 42.5 | |
| CLIPSelfBackbone=EVA02 ViT-L/14, Training Data=GTA-7+CoCo2026.02 | 51.2 | 59.7 | 40.6 | 39.7 | 48 | |
| SANBackbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 50.8 | 55.9 | 39.1 | 32.9 | 44.7 | |
| MAFT+Backbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 50.7 | 56.5 | 37.6 | 36.7 | 46.4 | |
| OVSegBackbone=EVA02 ViT-L/14, Training Data=GTA-72026.02 | 50.5 | 54.9 | 38.1 | 35.6 | 44.8 | |
| RSC-CLIPSelfBackbone=EVA02 ViT-B/16, Training Data=CS-7+COCO2026.02 | 50 | 53 | 40.1 | 35.1 | 44.5 | |
| CLIPSelfBackbone=EVA02 ViT-B/16, Training Data=CS-7+COCO2026.02 | 49.4 | 54.3 | 39.6 | 36.5 | 45 | |
| MaskAdapterBackbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 48.3 | 51.2 | 39.8 | 36.1 | 43.8 | |
| CAT-Seg+AdvStyleBackbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 48 | 47.9 | 36 | 36.5 | 42.1 | |
| CAT-SegBackbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 47.6 | 51.6 | 38.2 | 36.5 | 43.5 | |
| ESC-NetBackbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 47.3 | 49.2 | 40.1 | 34.5 | 42.7 | |
| MAFT+Backbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 46.4 | 48.8 | 38.1 | 35.4 | 42.2 | |
| CAT-Seg+DGInStyleBackbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 46.1 | 53.7 | 36.8 | 36 | 43.2 | |
| SANBackbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 44.3 | 46.8 | 35.4 | 30.7 | 39.3 | |
| OVSegBackbone=EVA02 ViT-B/16, Training Data=CS-72026.02 | 41.2 | 48.3 | 31.7 | 32.6 | 38.5 | |
| ProxyCLIPBackbone=EVA02+SAM ViT-L, Training Data=Training-Free2026.02 | 38.5 | 52.2 | 33.5 | 27.2 | 37.9 | |
| ProxyCLIPBackbone=CLIP+SAM ViT-L/14, Training Data=Training-Free2026.02 | 38.5 | 52.2 | 33.5 | 27.2 | 37.9 | |
| DeCLIPBackbone=EVA02+DINOV2 ViT-L, Training Data=COCO2026.02 | 34.2 | 47 | 32 | 28.3 | 35.4 | |
| ClearCLIPBackbone=EVA02 ViT-L, Training Data=Training-Free2026.02 | 33.5 | 34 | 26.6 | 21.6 | 28.9 | |
| ClearCLIPBackbone=CLIP ViT-L/14, Training Data=Training-Free2026.02 | 33.5 | 34 | 26.6 | 21.6 | 28.9 | |
| CAT-SegBackbone=EVA02 ViT-L/14, Training Data=COCO2026.02 | 33.2 | 42.8 | 28.6 | 27.6 | 33.1 | |
| CLIP-DINOiserBackbone=EVA02+DINOV2 ViT-L, Training Data=Training-Free2026.02 | 31.9 | 34 | 26.9 | 21.6 | 28.6 | |
| CLIP-DINOiserBackbone=CLIP+DinoV2 ViT-L/14, Training Data=Training-Free2026.02 | 31.9 | 34 | 26.9 | 21.6 | 28.6 |