Diffuse to Choose is a diffusion-based image repair model primarily designed for virtual try-on scenarios. It can preserve the details of the reference item while repairing the image and perform accurate semantic operations. By directly integrating the detail feature of the reference image into the latent feature map of the main diffusion model and combining it with perceptual loss to further preserve the details of the reference item, the model achieves a good balance between fast inference and high-fidelity details.