Multi-image Visual Question Answering on MMXU (test)
28.9Worsen RateMed-Flamingo
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Med-FlamingoSize=7B2025.02 | 28.9 | 30.4 | 25.1 | 28.1 | |
| Llava-medSize=7B2025.02 | 29.3 | 33.4 | 32.3 | 31.7 | |
| HuatuoGPT-VisionSize=7B2025.02 | 62.9 | 50.2 | 35.4 | 49.5 |