imerkle opened this issue 7 years ago · 1 comments
Sorry, all training images are facial images, thus cloth details are hardly captured by our training images. Currently, we don't have a plan to improve this part.