Oxygen-TryOn: Fashion-Native Foundation Model for Any-item Virtual Try-On

Researchers unveiled Oxygen-TryOn, a fashion-native virtual try-on foundation model.

It was trained with a dedicated data engine and a three-stage pipeline: continued pre-training, supervised fine-tuning, and reinforcement learning.

The model handles many fashion categories, multiple reference images, and multi-item outfits while preserving identity and item details.

It reports state-of-the-art results on both public and internal benchmarks.