Yolov11x segmentation model trained on images of 2d penises. This model was trained on humanoid and equine and canine penises, some more exotic ones are also in the data set.
Trained using Ultralytics 8.3.135. You may need to update for compatibility
You can find the realistic version here https://civarchive.com/models/301092
Performance
Version 2 offers 8% better Accuracy over v1 and 13.3% file size decrease.
How to use in ComfyUI
Install Impact Pack and Subpack
Install ComfyUI-Manager from https://github.com/ltdrdata/ComfyUI-Manager
Open the Manger window
Open Install Custom Nodes
Search for and install ComfyUI Impact Pack
Search for and install ComfyUI Impact Subpack
Save the model to ~\ComfyUI\models\ultralytics\segm\
Node setup
Add a FaceDetailer node from the Impact Pack
Add a UltralyticsDetectorProvider node and Select the ADetailer model, connect BBOX_DETECTOR to bbox_detector, and SEGM_DETECTOR to segm_detector_opt
Description
Added Canines to the dataset
Trained on 606 examples
FAQ
Comments (4)
So, it's probably fine, since Civitai scanned it, but I thought I'd let you know, mmaitre314 picklescan says that the "import '__builtin__ getattr'" function is potentially dangerous, as it can allow the importing of dangerous functions.
The model was made using Ultralytics 8.1.47 installed with Conda (version 8.2.x was not used due to models being trained with it not being compatible with ComfyUI Impact Pack).
The Command used to make the model was `yolo train data="./anime.yml" project="2D" model="./yolov8x-seg.pt" cache="ram"`
Looking around online this import seems to affect every Yolov8 segmentation model.
kinda shame both dick adetailers on the site has no dataset mixed in for penetration.pretty much doesnt work even for stuff like masturbation or BJs
That is the goal for the next version. Some work was started on it before Software performance became the limiting factor. That has since been fixed.