tcy6/VisRAG_Pipeline · Apply for community grant: Academic project (ZeroGPU)

Dear Hugging Face Team,

I hope this message finds you well. I’m excited to share that our paper, VisRAG: Vision-based Retrieval-Augmented Generation, has introduced the first multi-modal RAG pipeline leveraging vision-language models (VLMs). Our approach achieves remarkable performance gains of 25–39% over traditional text-based RAG systems, highlighting the potential of integrating visual context into retrieval-augmented generation.

We are honored that this work has been accepted for presentation at ICLR 2025, further underscoring its contribution to the field.

To accelerate the development and deployment of our demo, we would like to apply for a ZeroGPU. This support would enable us to run the system more efficiently, making it more accessible for the community to explore and build upon.

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

Thank you very much for considering our application. I’d be more than happy to provide any additional information about our research.

Best regards,
Chaoyue Tang