Paper 2604.19748
Tstars-Tryon 1.0: Robust and Realistic Virtual Try-On for Diverse Fashion Items
- Published
- Apr 2026
- Research lab
- Independent
- Citations
- 2
- GitHub
- Not linked
01 In brief
Summary
Tstars-Tryon 1.0, developed by the Pailitao Team at Alibaba Group, is a commercial-scale virtual try-on system designed for robustness, realism, versatility, and efficiency.
It handles challenging real-world cases like extreme poses, lighting variations, and motion blur, while preserving garment details and avoiding synthetic artifacts.
The system supports multi-image composition (up to 6 reference images) across 8 fashion categories (tops, pants, skirts, dresses, coats, shoes, bags, hats) with control over identity and background.
It uses a unified MMDiT architecture, a scalable data engine, and multi-stage training including reinforcement learning and distillation, achieving 3.92s for single-garment and 6.74s for multi-garment try-on.
The Tstars-VTON Benchmark, with 1780 paired samples, evaluates performance across four dimensions: Identity Consistency, Garment Fidelity, Background Preservation, and Physical/Structural Logic.
Quantitative results show Tstars-Tryon 1.0 outperforms top models like GPT-Image-2 and Nano Banana Pro, especially in multi-garment scenarios.
Human evaluations confirm its superiority, with win rates increasing with task complexity.
The system is deployed on the Taobao App, serving millions of users with tens of millions of try-on requests, and demonstrates strong zero-shot generalization on academic benchmarks like VITON-HD and DressCode.
02 From the paper
Abstract
Recent advances in image generation and editing have opened new opportunities for virtual try-on. However, existing methods still struggle to meet complex real-world demands. We present Tstars-Tryon 1.0, a commercial-scale virtual try-on system that is robust, realistic, versatile, and highly efficient. First, our system maintains a high success rate across challenging cases like extreme poses, severe illumination variations, motion blur, and other in-the-wild conditions. Second, it delivers highly photorealistic results with fine-grained details, faithfully preserving garment texture, material properties, and structural characteristics, while largely avoiding common AI-generated artifacts. Third, beyond apparel try-on, our model supports flexible multi-image composition (up to 6 reference images) across 8 fashion categories, with coordinated control over person identity and background. Fourth, to overcome the latency bottlenecks of commercial deployment, our system is heavily optimized for inference speed, delivering near real-time generation for a seamless user experience. These capabilities are enabled by an integrated system design spanning end-to-end model architecture, a scalable data engine, robust infrastructure, and a multi-stage training paradigm. Extensive evaluation and large-scale product deployment demonstrate that Tstars-Tryon1.0 achieves leading overall performance. To support future research, we also release a comprehensive benchmark. The model has been deployed at an industrial scale on the Taobao App, serving millions of users with tens of millions of requests.