馃拑 Wan-Dancer-14B

A Hierarchical Framework for Minute-scale Coherent Music-to-Dance Generation. Generate long-duration, high-quality, rhythmic dance videos from music with global structure and temporal continuity.

馃 Model 馃搼 Paper 馃寪 Project 馃捇 GitHub
鈿狅笍 Interactive demo note: The full music-to-dance generation runs on a 14B model requiring a GPU. Hosting the interactive Gradio demo needs a Hugging Face PRO account (or a 30-day-old account / community grant). Below are example outputs generated by Wan-Dancer-14B across dance styles.

Example Generations

Chinese Classical Dance

Chinese Classical Dance

seed=0 路 48 steps
Street Dance

Street Dance

seed=0 路 48 steps
K-Pop Dance

K-Pop Dance

seed=0 路 48 steps
Latin Dance

Latin Dance

seed=0 路 48 steps
Tap Dance

Tap Dance

seed=0 路 48 steps

How to run the interactive demo

Clone the repo and run locally with a CUDA GPU:

pip install -r requirements.txt
python app.py

The Gradio app takes a reference image and a dance prompt and returns a generated dance video using Wan-AI/Wan-Dancer-14B from Diffusers.