！日本語の説明は下部にあります！
This model detects NSFW areas (segments) in anime images. One model supports multiple types of detection.
Detectable Regions: nipples, pussy, anus, penis, testicles, x-ray, cross-section
Recommended Threshold: 0.5 ~ 0.8
Base Model: yolov11s-seg.pt
🗂️Install Location
Stable Diffusion WebUI/reForge -> models/adetailer
ComfyUI -> models/ultralytics/segm
❓FAQ
How to Deal with the Issue of Detected Areas Being Replaced by Another Object
If the detailer prompt is empty, enter the target prompt.
For example, if a nipples is being replaced with pussy, enter "nipples" in the prompt.
Lower the denoise value.
Expand the inpainting region.
WebUI (ADetailer): Increase "Mask erosion (-) / dilation (+)"
If "Inpaint only masked" is enabled, increase "Inpaint only masked padding, pixels"
ComfyUI (SEGS): Increase "crop_factor"
Troubleshooting “UltralyticsDetectorProvider Weights only load failed…” in ComfyUI
⚠️This workaround is required for v4 and earlier
Install the latest ComfyUI-Impact-Subpack and restart ComfyUI.
Open the file
<COMFYUI_INSTALL_PATH>\user\default\ComfyUI-Impact-Subpack\model-whitelist.txtin a text editor (e.g. Notepad).Add the following line to the end of the file, then save:
ntd11_anime_nsfw_segm_v4_all.pt
ntd11_anime_nsfw_segm_v4_anus.pt
ntd11_anime_nsfw_segm_v4_cross-section.pt
ntd11_anime_nsfw_segm_v4_nipples.pt
ntd11_anime_nsfw_segm_v4_penis.pt
ntd11_anime_nsfw_segm_v4_pussy.pt
ntd11_anime_nsfw_segm_v4_testicles.pt
ntd11_anime_nsfw_segm_v4_x-ray.ptRestart ComfyUI.
References
ComfyUI-Impact-Subpack README, Pull request
📄Note
ComfyUI allows you to filter detection class while using the "All" model.
Trained on images generated by Illustrious/noob derived models such as WAI-NSFW-illustrious-SDXL, Hassaku XL (Illustrious), Raehoshi illust XL, paruparu
Cannot be detected in monochrome images or comics
Cannot be detected in realistic images
⚙Training Settings (v5)
task="segment"
epochs=200
lr0=0.02
lrf=0.05
seed=0
imgsz=1024
batch=-1
close_mosaic=0
mosaic=0.0
erasing=0.0
scale=0.0For description of each setting, see: https://docs.ultralytics.com/modes/train/#train-settings
📢Past Announcements
Announcement for Stable Diffusion WebUI / reForge Users (2025/06/27)
I have an announcement for those using Stable Diffusion WebUI or reForge.
If you’re interested, please check out the details below—and I’d really appreciate any feedback or bug reports you may have!
When using my detection model with Stable Diffusion WebUI or reForge, the typical approach is to use it with the ADetailer extension.
While ADetailer offers many features, a few key functions were missing for making full use of my model.
To address that, I forked the original ADetailer and added the following features:
Class name filter: You can now filter detection results by specific class names
Class-specific prompts support: You can assign different prompts for each detected class using the
[CLASS=name]syntaxBounding box mask option: You can choose to use the bounding box as the mask for segmentation models, instead of the segmentation mask itself
You can find the repository here: https://github.com/newtextdoc1111/adetailer
Please see the README for installation instructions and more details on the new features.
My hope is that with class filtering, there will be no need to use anything other than the “ALL” model, and that the [CLASS=name] syntax will help reduce cases where unintended objects are drawn during inpainting.
Honestly, not having to train individual models for each release in the future would be a huge relief—this is one of the main reasons I implemented these changes 😂
If you’re someone who would still prefer to have individual models in future releases, feel free to share your specific use case. I may be able to offer an alternative solution (ComfyUI users are welcome too!).
アニメ画像のNSFW領域(セグメント)を検出するモデルです。一つのモデルで複数種類の検出に対応しています。
検出可能な領域： nipples, pussy, anus, penis, testicles, x-ray, cross-section
推奨閾値: 0.5 ~ 0.8
ベースモデル: yolov11s-seg.pt
🗂️インストール場所
Stable Diffusion WebUI/reForge -> models/adetailer
ComfyUI -> models/ultralytics/segm
❓FAQ
検出された領域が別のオブジェクトに描き換えられてしまう問題の対処法について
Detailer用プロンプトが空の場合、対象のプロンプトを入力する
nipplesがpussyに描き変わる場合、プロンプトに"nipples"と入力する
Denoseを下げる
インペイント領域を拡張する
WebUI (ADetailer): "Mask erosion (-) / dilation (+)" を増やす
ComfyUI (SEGS): "crop_factor" を増やす
ComfyUIで生成実行時に「UltralyticsDetectorProvider Weights only load failed. ...」と表示される問題の対処法について
⚠️v4以前のモデルで必要な対処法です
最新の ComfyUI-Impact-Subpack をインストールし、ComfyUIを再起動
<COMFYUIインストールパス>\user\default\ComfyUI-Impact-Subpack\model-whitelist.txtが作成されているので、メモ帳などで開く以下のテキストを追記して保存する
ntd11_anime_nsfw_segm_v4_all.pt
ntd11_anime_nsfw_segm_v4_anus.pt
ntd11_anime_nsfw_segm_v4_cross-section.pt
ntd11_anime_nsfw_segm_v4_nipples.pt
ntd11_anime_nsfw_segm_v4_penis.pt
ntd11_anime_nsfw_segm_v4_pussy.pt
ntd11_anime_nsfw_segm_v4_testicles.pt
ntd11_anime_nsfw_segm_v4_x-ray.ptComfyUIを再起動する
参考資料
📄メモ
ComfyUIの場合、"All" モデルを使いつつ検出したい種類をフィルタリング可能です
WAI-NSFW-illustrious-SDXL, Hassaku XL (Illustrious), Raehoshi illust XL, paruparu等のIllustrious/noob派生モデルで生成した画像で学習しています
モノクロ画像やコミックでは検出できません
リアル調の画像では検出できません
⚙学習設定 (v5)
task="segment"
epochs=200
lr0=0.02
lrf=0.05
seed=0
imgsz=1024
batch=-1
close_mosaic=0
mosaic=0.0
erasing=0.0
scale=0.0各設定の説明についてはこちら： https://docs.ultralytics.com/modes/train/#train-settings
📢過去のお知らせ
Stable Diffusion WebUI / reForge をお使いの方へのお知らせ（2025/06/27）
Stable Diffusion WebUI / reForge をお使いの方にお知らせがあります。
気になった方は以下の内容を確認し、感想や不具合の報告を頂けると嬉しいです！
Stable Diffusion WebUI / reForge （以下WebUIに統一）で私の検出モデルを利用する場合、ADetailerの拡張機能を使う方法が一般的です。
ADetailerは多くの機能がありますが、私の検出モデルを便利に活用するための機能がいくつか不足していました。
そこで、元のADetailerをフォークし以下の機能を追加しました。
クラスフィルター：検出したい対象をクラス名でフィルタリングする機能
クラス毎のプロンプト指定：
[CLASS=name]の構文で検出したクラス毎に個別のプロンプトを指定する機能バウンディングマスクオプション：セグメント検出モデルでもバウディングボックス形状でマスクを生成するためのオプション
リポジトリは https://github.com/newtextdoc1111/adetailer にあります。追加機能の詳細とインストール方法はREADMEを確認してください。
私の想定では、クラスフィルターがあれば「ALL」モデル以外不要になり、 [CLASS=name] 構文で適切なプロンプトを与えればインペイント後に別のオブジェクトが描画される事故も減るはずです。
特に、今後のリリースで複数のモデルファイルを学習しなくて済むと本当に助かります。このために実装したと言っても過言ではありません😂
今後のリリースでも個別モデルを残してもらいたい、と考えている方は具体的なユースケースを教えてもらえれば別の解決策を提案できるかもしれません。（ComfyUIユーザーの方からでも大丈夫です）。
Description
Initial release.
FAQ
Comments (19)
this can not detect spread pussy
Does lowering the threshold still not detect it?
Although the number of images is small, I have included it in the dataset. If you could post the image, I might be able to identify the cause.
@newtextdoc11996 can confirm. Even at 0.1 detection it cant detect spread pussy. Tried it on multiple angles.
This actually works much better than i would expect. Would love to see if you add spread pussyand covered nipples(Maybe cameltoe too) to the things it can detect too.
Good:
-Penis,Nipples,Anus and Pussy detections are pretty good
Bad:
-Can't Detect Spread Pussy
-Cross section false flags hair at even 0.5 detection sometimes
- Nipples being in All in 1 version causes issues cuz even at 0.3 denoising it can turn nipples into pussies.
Also i would say you should completely seperate nipples part from the rest in All in one version. You cant really go above 0.35 denoising because of nipples turning into pussies.
Thank you for your feedback and opinions. I'm glad you liked it.
I've received multiple comments about spread pussy not being detected, so I'm planning to improve this in version 2.0 (release date TBD released).
For now, I’ll wait to see if others have similar feedback before making any changes regarding these points.
- Adding detection for covered nipples and cameltoe
- Improving the detection rate for cross-section, x-ray
> Nipples being in All in 1 version causes issues cuz even at 0.3 denoising it can turn nipples into pussies.
> Also i would say you should completely seperate nipples part from the rest in All in one version. You cant really go above 0.35 denoising because of nipples turning into pussies.
The Ultralytics detection model only detects regions and does not affect the drawing behavior of Detailer. To prevent objects from being replaced by another after the Detailer process, try the following:
- If the Detailer prompt field is empty, enter the target object instead (in this case, "nipples").
- Expand the detected region:
-- WebUI (ADetailer): Increase "Mask dilation."
-- ComfyUI (SEGS): Increase "crop_factor."
Note: If you're using ComfyUI (SEGS), you can filter and input prompts for each class after detection.
@newtextdoc11996 I am using Webui so adding "nipples" to prompt wouldnt make sense because it would apply to both nipples,pussy, penis, etc and ruin it for the rest. I will try mask dilation tho.
Edit: Thanks,Dilation helped. Can go up to 0.4 denoising now without nipples turning into pussies.
My biggest problem right now is hair pieces getting false flagged as cross section now and if anyone else reads this and having the same issue it doesnt occur above 0.5 detection threshold. You will lose some detection on penis cuz those tend to need lower detection threshold but worth the tradeoff for now
@Wizzi Glad to hear it has improved!
How about using the individual detection versions for the nipples, pussy, and anus instead of the all-in-one version?
ADetailer allows multiple detection models to be used in a single image generation. These are the tabs labeled 1st, 2nd, 3rd, 4th (by default, four tabs should be visible).
If you need more tabs, go to Settings -> ADetailer -> Max tabs, increase the number, and then reload the UI.
Additional note:
ADetailer does not have a built-in function to filter detected regions by class name in the all-in-one version, so I recommend using the individual versions.
As for individual versions for the penis, x-ray, cross-section, I will consider creating them if there are requests from other users.
@newtextdoc11996 Yeah i know but this is the best penis adetailer in the site right now. Only one that can detect penis during sex scenes. Only one i have a problem with right now is cross section, x-ray doesnt have much false flagging and the rest works pretty well especially if you can lower the detection to 0.1-0.2(but you cant do that cuz you get 10 cross-section detections below 0.2 from bunch of hair strands fsr lol)
And honestly i like it all being in one adetailer. Currently i am running the all in one first at low weight and then applying seperate ones with personalized prompts which works perfectly.
@Wizzi I’ve released v2.0 with improved spread pussy detection rate.
Let me know if you don't see any improvements!
Just curious, what tool did you use for the image annotation? I'm trying to to fine-tune my own detection model for my own use case, but not sure where to starts. If you can give me a bit of your wisdom that you be greatly appreciated 🙏
Hi, I'm using CVAT on Windows. It's a free tool that can be installed locally, so you can handle datasets safely.
However, if you encounter any issues, you might need some knowledge of Git, Docker, Linux, and Python.
You can check the installation guide at the following URL:
https://docs.cvat.ai/docs/administration/basics/installation/
Another tool you can use locally is LabelMe, though I haven't used it myself.
@newtextdoc11996 Thank you for the quick respond! I was looking at CVAT myself but I don't know you can use it locally. Another quick question if you don't mind. How many images did you use to fine-tune this model?
@stableskynet171 I'm using approximately 600 images for training. These images were selected from those originally stored locally.
However, I didn't start with all 600 images from the beginning. Instead, I followed an iterative process: annotating 100 images at a time, training the model, testing it, and repeating the cycle until I achieved satisfactory results— which happened at around 600 images.
If you have any other questions, I'll be happy to answer them as much as I can.
@newtextdoc11996 Thank you for the reply! I'm going to try and annotate some images over the weekend. If anything else come up, I'll sure to let you know, thank you :D
testiclesを検出しないモデルを探していたのでpenis単体があったら嬉しいです！🙇
penis単体（penisだけを検出するモデル）を作成するのは難しくないので、version 2.0で追加しておきます。
@mamiyarinzou penisだけを検出する個別のモデルを追加したv2.0をリリースしたのでお知らせします 📢
@newtextdoc11996 ありがとうございます！🙇