README.md

---
title: FloodDiffusion Streaming Demo
emoji: "\U0001F30A"
colorFrom: blue
colorTo: purple
sdk: docker
app_port: 7860
hardware: t4-small
pinned: true
license: apache-2.0
short_description: "Streaming Motion Generation (CVPR 2026)"
---

# FloodDiffusion: Tailored Diffusion Forcing for Streaming Motion Generation

**CVPR 2026**

Yiyi Cai, Yuhan Wu, Kunhang Li, You Zhou, Bo Zheng, [Haiyang Liu](https://h-liu1997.github.io/)

[![arXiv](https://img.shields.io/badge/arXiv-2512.03520-blue)](https://arxiv.org/abs/2512.03520)
[![Project Page](https://img.shields.io/badge/Project_Page-FloodDiffusion-orange)](https://shandaai.github.io/FloodDiffusion/)
[![Code](https://img.shields.io/badge/Github-Code-green)](https://github.com/ShandaAI/FloodDiffusion)
[![Model](https://img.shields.io/badge/HF-Model-yellow)](https://huggingface.co/ShandaAI/FloodDiffusionTiny)

Interactive streaming demo for real-time, infinite-length, text-driven 3D human motion generation. Type a motion description, click Start, and watch the skeleton move in real-time. Change text on-the-fly to seamlessly transition between motions.