Visual Question Answering on CV-Bench
88.9AccuracySR-3D
Evaluation results
| Method | Links | |
|---|---|---|
| SR-3DModel Version=base, Inference Mode=base2026.06 | 88.9 | |
| SR-ReaL-LORModel Version=Full, Inference Mode=LOR2026.06 | 88.5 | |
| SR-ReaL-directModel Version=Full, Inference Mode=direct2026.06 | 88.1 | |
| SeProD2026.05 | 78.4 | |
| Mini-o32026.05 | 74.4 | |
| Qwen2.5-VL2026.05 | 73.9 | |
| LLaVA-1.5-13BModel=LLaVA-1.5-13B, Cropped Images=false2025.05 | 61.6 | |
| LLaVA-1.5-13B + QLIPModel=LLaVA-1.5-13B, Cropped Images=true2025.05 | 60.7 | |
| LLaVA-1.5-7b + QLIPModel=LLaVA-1.5-7b, Cropped Images=false2025.05 | 40.2 | |
| LLaVA-1.5-7bModel=LLaVA-1.5-7b, Cropped Images=false2025.05 | 39.9 |