First, download the sdxl and sdxl inpainting models into the pretrained_models folder using Git LFS. Human parsing model and our checkpoint on VITONHD should be ...