Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Official Pytorch Code of the Paper "FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization"
| Date | Stars |
|---|---|
| 2026-07-24 | 251 |
| 2026-07-25 | 251 |
| 2026-07-28 | 251 |
| 2026-07-30 | 251 |
| 2026-08-06 | 251 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
<div align="center">
<h1>
<img src="assets/fashionchameleon.png" width="40" style="vertical-align: middle;" />
FashionChameleon: Towards Real-Time and Interactive Human-Garment Video Customization
</h1>
<p align="center">
<span>
<a href="https://arxiv.org/pdf/2605.15824" target="_blank">
<img src='https://img.shields.io/badge/2605.15824-FashionChameleon-red' alt='Paper PDF'></a>    
</span>
<span>
<a href='https://quanjiansong.github.io/projects/FashionChameleon' target="_blank">
<img src='https://img.shields.io/badge/Project_Page-FashionChameleon-green' alt='Project Page'></a>    
</span>
<br>
<span>
<a href='https://huggingface.co/papers/2605.15824' target="_blank">
<img src='https://img.shields.io/badge/Hugging_Face-FashionChameleon-blue' alt='Hugging Face'></a>    
</span>
<span>
<a href='https://huggingface.co/datasets/QuanjianSong/HGC-Bench' target="_blank">
<img src='https://img.shields.io/badge/Hugging_Face-HGC--Bench-yellow' alt='HGC-Bench'></a>    
</span>
</p>
<br/>
<div align="center">
<b>TL;DR:</b><br/>
We propose <span className="text-white font-medium">FashionChameleon</span>, a real-time and interactive framework for human-garment customization in streaming autoregressive video generation.
It achieves real-time generation at 23.8 FPS on a single GPU.
</div>
<img src="assets/teaser.png" style="width:100%; height:100%;"/>
</div>
## 📅 Todo
- [ ] Release the checkpoint.
- [ ] Release the training-free kv cache rescheduling for interactive inference.
- [x] 🔥 Release the code (Wan2.2-TI2V-5B) for gradient-reweighted dmd and the corresponding inference.
- [x] 🔥 Release the code (Wan2.2-TI2V-5B) for in-context teacher forcing and the corresponding inference.
- [x] 🔥 Release the code (Wan2.2-TI2V-5B) for in-context sft and the corresponding inference.
- [x] 🔥 Release the <a href="https://huggingface.co/datasets/QuanjianSong/HGC-Bench" target="_blank">HGC-Bench</a>.
- [x] 🔥 Release the <a href="https://quanjiansong.github.io/projects/FashionChameleon" target="_blank">Project Page</a>.
- [x] 🔥 Release the <a href="https://arxiv.org/pdf/2605.15824" target="_blank">Technical Report</a>.
## ✨ Highlight
> **1. Interactive Customization.** We train a single-garment switching teacher using tailored I2V priors and mismatched reference–garment pairs. During generation, we introduce KV-cache rescheduling to enable interactive multi-garment customization without requiring video data containing multi-garment switching.
> **2. Gradient-Reweighted DMD.** Traditional self-forcing treats all self-rolled frames equally during DMD backpropagation. However, later frames typically suffer from larger quality degradation and thus require stronger gradient supervision. We dynamically reweight DMD gradients during self-rolling using a reward model to improve extrapolation consistency.
> **3. Real-Time Generation.** Through streaming distillation with in-context learning, FashionChameleon achieves 23.8 FPS for 720p generation on a single H200 GPU, 30–180× faster than existing customization methods.
<img src="assets/intro.png" style="width:100%; height:100%;"/>
## 🎬 Overview
***FashionChameleon*** is built upon [Wan2.2-TI2V-5B](https://huggingface.co/Wan-AI/Wan2.2-TI2V-5B), featuring: **(i)** Teacher Model with In-Context Learning, **(ii)** Streaming Distillation with In-Context Learning, and **(iii)** Training-Free KV Cache Rescheduling.
<img src="assets/overall_framework.png" style="width:100%; height:100%;"/>
## 🔧 Step0. Setup
### Prepare Environment
```
git clone https://github.com/QuanjianSong/FashionChameleon.git
cd FashionChameleon
# Installation with the requirement.txt
conda create -n FashionChameleon python=3.10
conda activate FashionChameleon
pip install -r requirements.txt
```
### Download Backbones
Our FashionChameleon is built upon [Wan2.2-TI2V-5B](https://huggingfExcerpt of 10,683 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:051ed235c3c9f71d, topic:sft