Introducing a strong VLM FIRE with a powerful feedback-refinement capability.

Created in August 10, 2024

2024   ·   VLM   ·   VLM

Can VLMs refine their responses based on user feedback spontaneously❓

Introducing 🔥FIRE: A Dataset for Feedback Integration and Refinement Evaluation of Multimodal Models.

  • 👉 Project: https://mm-fire.github.io

Checkout our 🚀 Gradio Demo

Feel free to leave a message to me, if you have issues about relevant demo and code.