sd-webui-controlnet
sd-webui-controlnet copied to clipboard
Support Face Landmark Control
It's very useful Generate face with the identical poses and expression with Face-Landmark-ControlNet
Is possible to support it ? Thank you very much
I've made a PR #637
Face landmarks will be officially supported by a official model of ControlNet V1.1 openpose-full model, trained with arbitary combinations of face, body, and hand landmarks. The annotator will be a pytorch version of openpose's full 3 models, written by ControlNet team. The community model and anotator might be very constructive but may be replace by official ones that will be released soon.
Face landmarks will be officially supported by a official model of ControlNet V1.1 openpose-full model, trained with arbitary combinations of face, body, and hand landmarks. The annotator will be a pytorch version of openpose's full 3 models, written by ControlNet team. The community model and anotator might be very constructive but may be replace by official ones that will be released soon.
Oh that's really great news! I have been able to get nice results using the ControlNet face-landmark model. Looking forward to the new ControlNet V 1.1!
another news: https://github.com/crucible-ai/ControlNet/blob/laion_dataset/README_laion_face.md https://huggingface.co/CrucibleAI/ControlNetMediaPipeFace