Semantic Segmentation on ADE20K (1 64)
47.9mIoUCLIP-Guided SAM
Evaluation Results
| Method | Links | |
|---|---|---|
| CLIP-Guided SAMFine-tuning SAM?=Yes, VLM Size=87M (CS-CLIP-Base)2026.05 | 47.9 | |
| GroundedSAMFine-tuning SAM?=No, VLM Size=172M (GroundingDINO)2026.05 | 42.4 | |
| UniMatch V2 (published)Encoder=DINOv2-B2026.07 | 38.7 | |
| SemiVLEncoder=CLIP-B2026.07 | 33.7 | |
| UniMatch V2 (published)Encoder=DINOv2-S2026.07 | 31.5 | |
| UniMatchEncoder=CLIP-B2026.07 | 25.3 | |
| CLIPSurgery+SAMFine-tuning SAM?=Yes (adapters + decoder), VLM Size=87M (CS-CLIP-Base)2026.05 | 24.8 | |
| CLIPSurgery+SAM (modified)Fine-tuning SAM?=No, VLM Size=87M (CS-CLIP-Base)2026.05 | 23.4 | |
| CLIPSurgery+SAMFine-tuning SAM?=Yes (decoder), VLM Size=87M (CS-CLIP-Base)2026.05 | 23.1 | |
| UniMatchEncoder=RN-1012026.07 | 21.6 | |
| CLIPSurgery+SAM (vanilla)Fine-tuning SAM?=No, VLM Size=87M (CS-CLIP-Base)2026.05 | 20 |