Spaces:

tcy6

/

VisRAG_Pipeline

like 12

Running

🚀 Our paper VisRAG: Vision-based Retrieval-Augmented Generation introduces the first multi-modal RAG pipeline using vision-language models (VLMs), achieving 25–39% performance gains over traditional text-based RAG systems! 🌟

To run the demo faster and more efficiently, we would like to apply for a GPU grant to support further development of VisRAG.

🔗 Paper link: https://arxiv.org/abs/2410.10594

I’m happy to answer any questions about our research. Thanks for your support!
—tcy6