---
license: apache-2.0
tags:
- flow-matching
- diffusion
- generative-ai
- biology
- chemistry
pipeline_tag: other
library_name: pytorch
---

## ⚡️ReQFlow: Rectified Quaternion Flow for Efficient and High-Quality Protein Backbone Generation
[![arXiv](https://img.shields.io/badge/arXiv-2502.14637-b31b1b?style=flat&logo=arxiv)](https://arxiv.org/abs/2502.14637)[![Github](https://img.shields.io/badge/Github-Code-blue?style=flat&logo=github)](https://github.com/AngxiaoYue/ReQFlow)[![Hugging Face](https://img.shields.io/badge/Hugging%20Face-Model-orange?style=flat&logo=huggingface)](https://huggingface.co/AngxiaoYue/ReQFlow/tree/main)


Our **ReQFlow** achieves **state-of-the-art (SOTA)** performance 🎯 in protein backbone generation while requiring significantly fewer sampling steps and substantially reducing inference time. For example, it is **37×** faster than RFDiffusion and **62×** faster than Genie2 when generating a backbone of length 300, demonstrating both its effectiveness and efficiency 🚀🚀🔥.
<p align="center" style="margin-bottom: -2.8px;">
  <img src="assets/scheme_reqflow.png" width="95%">
</p>
<p align="center" style="margin-top: 1.5px;">
  <img src="assets/prot_vis.png" width="95%">
</p>
Ready to dive in? We’ve got checkpoints and step-by-step guides to get you inferencing in no time—playtime starts now! ✨

## 🚀🚀🔥
🔗arxiv link: [https://arxiv.org/abs/2502.14637](https://arxiv.org/abs/2502.14637)

🌳github link: [https://github.com/AngxiaoYue/ReQFlow](https://github.com/AngxiaoYue/ReQFlow)