ATTENTION: IF YOU HAVE DOWNLOADED MODELS BEFORE ALREADY - RE-DOWNLOAD THEM PLEASE. I SHOULD HAVE UPDATED THEM WITH NO DILL VERSION(EXCEPT MODEL THAT WAS CHOSEN AS NEW VERSION UPDATE, SO YOU CAN RECEIVE NOTIFICATIONS.
If you are having issues:
Activate your venv
pip install dillpip install --upgrade ultralytics
This should fix generic problems.
Small segmentation model aiming to help with head inpainting/enhance.
Note: It is not very accurate, and will struggle with long hair.
This version is smallest available, compromise in accuracy, but it's just 6mb. Medium(50mb) brings in up a notch, but don't think i will be releasing that one in particular.
Oke, medium is up.
Here are validation results on ~470 images:

Can also be downloaded at HF(whenever i can actually log in): https://huggingface.co/Anzhc/Anzhcs_YOLOs
Script im using: https://github.com/Anzhc/Training-script-for-Ultralytics-YOLO
Usage
Put into `automatic1111/models/adetailer`, or any other place that you know will work with YOLOv8 models.
Use.
What it does and what is it for?
Head segmentation, duh.

It tries to select head, including hair, while avoiding rest. It is hard to avoid rest...
Model is very inconsistent with long hair.
Overall it should be able to work with wide variety of styles and mediums, as i tries to create a robust enough dataset with quite a bit of abstract, traditional, real and digital content. It is not a big one though, so there is a large room for improvement.

But what you might need it for? To have some fun, while reducing tedium a bit.


P.S. Those inpaints were done on older version, which had a bit worse masking.
All inpaints were done directly with adetailer in single pass with inpaint model.
And it took a few tries, hard to get 7 inpaints right in a row xd
Description
THIS IS SAME MODEL AS MEDIUM V1 BUT WITHOUT DILL.
NEW VERSION TO SEND NOTIFICATIONS.
FAQ
Comments (14)
:NomNomNom:
Really good! I think it works great
For some reason, the file keep corrupting when it is used. Do you have any advice?
It's a great tool that even includes hair, but unfortunately I wasn't satisfied with the quality of the hair in the 6MB version. I look forward to the release of the 50MB version.
Similar issues to user JustTookLazy. I am using the web forged version and it doesn't work when I reinstall sd forged. Any solution?
not working with forge
No module named dill
Fantastic! Works like a charm. :)
can't use it in forge even after the 2 pip installs. Its like it doesn't recognize the file even if other adetaliers from civitai are recognized.
what do you mean by activate venv?
any chance you can make a adetailer model for the suspended thigh high?
I would love a male and female specific version of this like your face adetailer models at some point.
This model is really great, I use it all the time, thanks. I hope you will make some refinements in the future, it's worth it :-)
Hello :) where do i put that in a ComfyUI installation i try to add it in a "adetailer" folder but it dont work. Have a good day
Looks like we don't have an active mirror for this file right now.
CivArchive is a community-maintained index — we catalog mirrors that volunteers upload to HuggingFace, torrents, and other public hosts. Looks like no one has uploaded a copy of this file yet.
Some files do get recovered over time through contributions. If you're looking for this one, feel free to ask in Discord, or help preserve it if you have a copy.