Outfit Anyone
Alibaba research model for ultra-realistic virtual try-on on any body.
About
Outfit Anyone is a virtual try-on model from Alibaba's Institute for Intelligent Computing, published on arXiv in July 2024. It uses a two-stream conditional diffusion model that processes the person and garment separately before fusing them, producing high-fidelity try-ons that handle pose changes and body shape variation.
The model also works on anime characters and integrates with Animate Anyone for motion video. A free Hugging Face demo lets users try garments on preset AI models; custom person photos are blocked for privacy, and the research code is open source on GitHub.
What it can do
- Two-stream diffusion virtual try-on
- Works on diverse body types and anime characters
- Animate Anyone integration for motion video
- Open-source research code
Setup
Try the free Hugging Face demo or clone the open-source research code from GitHub.
Prerequisites
Garment images; a GPU environment for running the code locally.
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 123; Trending score 66/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.