OOTDiffusion is a Python-based implementation for controllable virtual try-on using Outfitting Fusion based Latent Diffusion. It supports model inference for both half-body and full-body scenarios and provides ONNX support for human parsing.
