Adetailer and face recognition error

[Ardogg]

When using adetailer, an error occurs related to the version of torch and torchvision. Previously, it was solved by the command:
pip install torch==2.1.0 torchvision==0.16.0

However, now adetailer simply does not transfer the face from the Lora model. Among other things, it also gives an error ip-adapter face.

I did not delve into the problem much, but perhaps all models on similar algorithms give an error. Faceswaplab works as usual.

The problem has been observed since the last update.

Also, even after entering the commands, no models are recognized except yolo_face

Maybe someone knows a solution?

[Ardogg]

Is that obsolete?