In the OFT paper, I noticed that you have used the ControlNet for face landmark2image , canny2image, densepose2image and segmentation2image. However, I actually only find the canny to image and segmentation2image pre-trained weights in the original ContorlNet gtihub repo. I wonder where should I refer to to obtain the fully fine-tuned ControlNet weights for other controllable image generation tasks?
Hi authors.
Sorry to bother you.
In the OFT paper, I noticed that you have used the ControlNet for face landmark2image , canny2image, densepose2image and segmentation2image. However, I actually only find the canny to image and segmentation2image pre-trained weights in the original ContorlNet gtihub repo. I wonder where should I refer to to obtain the fully fine-tuned ControlNet weights for other controllable image generation tasks?
Thanks.