ResearchBenchmarksVisual Question Answering on Yo’LLaVA Single-ConceptVisual Question Answering on Yo’LLaVA Single-ConceptFollow95.9VQA AccuracyPeKit89.6691.2892.994.52Feb 4, 2025Evaluation resultsMethodMethodLinksVQA AccuracyPeKitBackbone=InternVLBackbone=InternVL2025.0295.9PeKitBackbone=LLaVABackbone=LLaVA2025.0293.4Yo’LLaVABackbone=LLaVABackbone=LLaVA2025.0292.9LLaVABackbone=LLaVABackbone=LLaVA2025.0289.9