Visual Grounding on Long-text (val)
77.86IoUE3AD
Evaluation Results
| Method | Links | |
|---|---|---|
| E3ADBackbone=VLM2025.12 | 77.86 | |
| VLTVGBackbone=ResNet-1012025.12 | 68.8 | |
| VL-BERTBackbone=ResNet-1012025.12 | 66.7 | |
| RSD-LXMERTBackbone=ResNet-1012025.12 | 65.8 | |
| TransVGBackbone=ResNet-1012025.12 | 65.45 | |
| UNINEXTBackbone=ResNet-502025.12 | 65.32 | |
| CAVGBackbone=ViT2025.12 | 64.36 | |
| CMRTBackbone=ResNet-1522025.12 | 64.25 | |
| Grounding-DINOBackbone=ViT2025.12 | 63.15 | |
| MDERTBackbone=ResNet-1012025.12 | 62.72 | |
| CMSVGBackbone=EfficientNet2025.12 | 62.21 | |
| AttnGrounderBackbone=ResNet-502025.12 | 57.25 | |
| Qwen3-VL-8BBackbone=VLM2025.12 | 50.13 | |
| Qwen2.5-VL-72BBackbone=VLM2025.12 | 49.83 | |
| Qwen2.5-VL-7BBackbone=VLM2025.12 | 41.92 |