This repository implements a human-centric latent diffusion model for fashion image editing. It allows designers to manipulate and visualize clothing on human models using multimodal conditioning. The approach builds on stable diffusion techniques and is evaluated on standard fashion datasets like VITON-HD and DressCode.