🧢 CAP-VTON

Clothing agnostic Pre-inpainting Virtual Try-ON

---

📖Paper -- 💾Code -- 🕹️Colab_Demo

Visualization

Installation

Create a Conda Python environment and install requirements.

It runs on Linux (Ubuntu) environment...!

conda create -n capvton python==3.10
conda activate capvton
cd CAP-VTON
pip install -r requirements.txt

"Run Start_CaP_VTON.ipynb"

Specifications

• Capacity: 34.3GB
• RAM: 24GB more
• GPU: (Tested: RTX4080 and A100)

Acknowledgement

This work was developed by extending Leffa.
We would like to acknowledge the contributions of the original authors.
For in-depth technical details, please see the Leffa Paper.

Citation

If you find CAP-VTON helpful for your research, please cite our work:

@misc{DevChoco_CAP-VTON_2025,
  author       = {Sehyun Kim, Jiwoo Lee, and Hye Jun Lee, from DevChoco},
  journal      = {arXiv preprint arxiv.org:2509.17654},
  title        = {CAP-VTON: Clothing agnostic Pre-inpainting Virtual Try-ON},
  year         = {2025},
  howpublished = {\url{https://github.com/DevChoco/CAP-VTON}},
  note         = {GitHub repository, MIT License},
}