Visual Grounding on ImageNet-1K (val)
0.9Alignment ScoreGT (Labels)
Evaluation Results
| Method | Links | |
|---|---|---|
| GT (Labels)W (Weights)=-, D (Data)=-, Backbone=RN502026.03 | 0.9 | |
| GT (Labels)W (Weights)=-, D (Data)=-, Backbone=ViT2026.03 | 0.9 | |
| CLIPDissectW (Weights)=✓, D (Data)=✓, Backbone=RN502026.03 | 0.85 | |
| GT (Descr.)W (Weights)=✗, D (Data)=✓, Backbone=RN502026.03 | 0.76 | |
| GT (Descr.)W (Weights)=✗, D (Data)=✓, Backbone=ViT2026.03 | 0.76 | |
| WWWW (Weights)=✓, D (Data)=✓, Backbone=RN502026.03 | 0.66 | |
| WWWW (Weights)=✓, D (Data)=✓, Backbone=ViT2026.03 | 0.63 | |
| UNBOXW (Weights)=✗, D (Data)=✗, Backbone=RN502026.03 | 0.56 | |
| UNBOXW (Weights)=✗, D (Data)=✗, Backbone=ViT2026.03 | 0.43 | |
| DEXTERW (Weights)=✓, D (Data)=✗, Backbone=RN502026.03 | 0.42 | |
| CLIPDissectW (Weights)=✓, D (Data)=✓, Backbone=ViT2026.03 | 0.36 | |
| DEXTERW (Weights)=✓, D (Data)=✗, Backbone=ViT2026.03 | 0.36 | |
| RandomW (Weights)=✗, D (Data)=✗, Backbone=RN502026.03 | 0.03 | |
| RandomW (Weights)=✗, D (Data)=✗, Backbone=ViT2026.03 | 0.03 |