Visual Grounding on Lisa Grounding
75.69AccuracyVisual Jigsaw
Evaluation Results
| Method | Links | |
|---|---|---|
| Visual Jigsaw2025.12 | 75.69 | |
| Rotation + CL + CARE2025.12 | 74.61 | |
| GRPO-CARE2025.12 | 74.55 | |
| Jigsaw2025.12 | 74 | |
| Vision-Zero2025.12 | 73.82 | |
| Jigsaw + CARE2025.12 | 73.72 | |
| VisualSphinx2025.12 | 73.28 | |
| Jigsaw + CL2025.12 | 73.28 | |
| Qwen 2.5 VL2025.12 | 72.86 | |
| Jigsaw + CL + CARE2025.12 | 72.62 | |
| Mix + CL + CARE2025.12 | 70.91 | |
| JigsawPrompting=CoT prompting, Backbone=Qwen-VL-2.5 3B2025.12 | 68.21 | |
| VLM-R1Prompting=CoT prompting, Backbone=Qwen-VL-2.5 3B2025.12 | 67.79 | |
| MixPrompting=CoT prompting, Backbone=Qwen-VL-2.5 3B2025.12 | 67.19 | |
| PatchFit + CL + CARE2025.12 | 64.5 | |
| ViCrit2025.12 | 63.2 | |
| Qwen 2.5 VLPrompting=CoT prompting, Backbone=Qwen-VL-2.5 3B2025.12 | 61.52 | |
| Jigsaw-R1Prompting=CoT prompting, Backbone=Qwen-VL-2.5 3B2025.12 | 61.28 |