Visual Question Answering on PathVQA (Open)
39Token RecallMMed-RAG
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MMed-RAGBackbone=LLaVA-Med2025.08 | 39 | — | |
| LLaVA-MedBackbone=Vicuna2025.08 | 39 | — | |
| LLaVA-MedBackbone=BioMedCLIP2025.08 | 39 | — | |
| LLaVA-MedBackbone=LLaVA2025.08 | 38 | — | |
| CLAREBackbone=Qwen2-vl2025.08 | 38 | 37 | |
| CLAREBackbone (Size)=Qwen2-vl (7B)2025.08 | 38 | — | |
| CLAREBackbone=Pixtral2025.08 | 36 | 35 | |
| CLAREBackbone (Size)=Pixtral (12B)2025.08 | 36 | — | |
| FT RAGBackbone (Size)=Qwen2-vl (7B)2025.08 | 33 | — | |
| CLARE + PDistBackbone (Size)=Qwen2-vl (7B)2025.08 | 32 | — | |
| CLARE + PDistBackbone (Size)=Pixtral (12B)2025.08 | 32 | — | |
| MMed-RAGBackbone (Size)=LLaVA (7B)2025.08 | 31 | — | |
| FT RAGBackbone (Size)=Pixtral (12B)2025.08 | 31 | — | |
| ReaderBackbone (Size)=Pixtral (12B)2025.08 | 26 | — | |
| ReaderBackbone (Size)=Qwen2-vl (7B)2025.08 | 25 | — | |
| BiomedGPT-B2025.08 | — | 28 |