lllyasviel / ControlNet

Let us control diffusion models!
Apache License 2.0
30.28k stars 2.72k forks source link

OpenPifpaf pretrained data #467

Open p0mad opened 1 year ago

p0mad commented 1 year ago

Hi, As mentioned in the paper https://arxiv.org/abs/2302.05543:

Human Pose (OpenPifPaf) We use learning-based pose estimation method [27] to “find” humans from internet using a simple rule: an image with human must have at least 30% of the key points of the whole body detected. We obtain 80k pose-image-caption pairs. Note that we directly use visualized pose images with human skeletons as training condition. The model is trained with 400 GPU-hours on Nvidia RTX 3090TI. The base model is Stable Diffusion 2.1. (See also Fig. 8.)

you have trained the SD2.1 on openpifpaf whole body and that is fascinating!

But i was wondering why there are no pretrained model or demo for openpifpof wholebody in the github page nor in the @huggingface !

Are you plan to release the model? No matter how accurate it is, we all wanted to use it. @lllyasviel @williamyang1991 @scarbain @eltociear @camenduru @sethupavan12 image

I also found some other people asking about it in issues but no answer! @anwoflow @LCorleone @Olwaro @huytuong010101 @Paludgus @BlueAccords @ninjasaid2k @sALTaccount @richard-schwab @richard-schwab @TheLukaDragar

Thanks Best regards

geroldmeisinger commented 1 year ago

all duplicates about "Release the openpifpaf model": https://github.com/lllyasviel/ControlNet/issues/109 https://github.com/lllyasviel/ControlNet/issues/201 https://github.com/lllyasviel/ControlNet/issues/331 https://github.com/lllyasviel/ControlNet/issues/467