Ever ran into a situation where normal YOLO models in ADetailer failed to find a face on your image? I did, and so i decided to train my own YOLO model!

Introducing an FFAD, it stands for Furry Face ADetailer. Trained on over 5000 anthro images, it includes popular species like canines to more unique like avians. The final results are good, i will not go into details but it can mark an anthro face with around 95% which is way more than with normal YOLO model.

I strictly recommend setting the confidence of this model to 0.6 if you're generating an anthro face, bare minimum is 0.5 tho.