！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
This is a variant model of v5.0.
This model is designed to detect penis within the cross-section and x-ray areas.
You can see the specific differences in the second image of the showcase.
Aside from that, there are no other changes.
----
このモデルはcross-section, x-rayの領域に含まれるpenisを検出します。
詳しい違いはショーケースの二枚目の画像で確認出来ます。それ以外の違いはありません。
FAQ
Comments (30)
For 5.0, occasionally the nipples are detected as vagina.
Thank you for the report.
If possible, could you share the image or the combination of model and prompt you used? I'd like to use it as a reference for improving the model in future updates.
On my end, I'll check my existing dataset to see if there are any annotation (labeling) errors.
If the detection threshold is set to a relatively low value (0.2 ~ 0.5), increasing it to around 0.8 may help reduce false detections.
@newtextdoc11996 Thank you! I'll try that and see if that helps.
Nothing special with the prompt fed to the detailer, just had "nsfw, explicit"
As for the models it was WAI Illustrious v15 for my checkpoint with ExpressiveH v2
@soundwaveshogun Thank you for the information!
Just to confirm, which of the following two cases best describes the situation you commented?
The way to address the issue differs depending on which one it is, so it would be very helpful if you could let me know.
1. The nipples were replaced with a pussy in the generated image
2. An incorrect label was shown during detection
By “label,” referring to the bounding box shown around the detected regions in the sample images I posted.
For ADetailer, you can save labeled images by enabling “Save mask previews” in the settings.
Also, I checked the new dataset I added for version 5.0.
There was only one missed annotation for another body part, so it seems unlikely to have a significant impact on detection accuracy.
フォークのadetailerの話もこちらで大丈夫でしょうか？
最新のultralyticsがadetailerと互換性がないということでultralyticsのバージョンを固定したフォーク（Byt3Sh0ck/adetailer）が出ています。newtextdoc1111版でも同様の処理が必要かもしれません。
https://www.reddit.com/r/StableDiffusion/comments/1oc3410/comment/nlxw5du/
情報共有ありがとうございます！
フォーク版を確認しましたが、Ultralyticsという依存ライブラリのバージョンを固定すれば解決するようです。
ただ、検証に時間がかかりそうなので確保できたときに対応しようと思います。
進捗があればこのコメント欄の返信やアナウンスとして皆さんに告知しようと思います。
@jiayue こんにちは。ultralyticsの互換性について調査し、原因を特定したので修正をフォーク版に反映しました。
私の手元では対象となるultralyticsのバージョンv8.3.217以前・以降どちらでも動作することを確認しています。
原因と解決方法の詳しい情報はオリジナルリポジトリのIssueにコメントしたので、良ければご確認ください。
https://github.com/Bing-su/adetailer/issues/817#issuecomment-3600170527
I'm having trouble with the stable diffusion reforge.
Detection with the Adetailer works, but the mask doesn't work.
I confirmed that inverting the mask area works for all areas, including the detected area.
I suspect there's a problem with setting the mask area after detection.
This issue occurred in both v4 and v5.
I did not install Adetailer additionally, but used the one included by default.
Ultralytics versions 8.3.221 and 8.3.231 failed to detect segments but bbox works.
downgrading Ultralytics to version 8.2.103 resolved the issue. now works for segments.
I'm using reforge.
Hello, and sorry for the delayed reply.
I tested my forked version of adetailer on my end (reforge), and it worked correctly with the latest Ultralytics version 8.3.233. It seems that certain Ultralytics versions were causing the issue.
I will gather a bit more information and decide whether I should lock the Ultralytics version used in my forked adetailer.
@newtextdoc11996 thank you. now it works great.
@fragbait Hi! After additional investigation, I've identified the cause of the issue.
I've applied the fix to my forked adetailer repository and confirmed that it works with both Ultralytics versions before and after v8.3.217.
For more detailed information, please check the comment I posted on the issue discussion in the original repository:
https://github.com/Bing-su/adetailer/issues/817#issuecomment-3600170527
@newtextdoc11996 How do I apply this? Do I just need to update ultralytics in the reforge?
@zzozz Yes. To check whether the issue is resolved with the newer Ultralytics, you will need to update Ultralytics and use the adetailer version I forked.
Installation instructions for the forked version are available in the README below:
https://github.com/newtextdoc1111/adetailer?tab=readme-ov-file#installation-fork-version
That said, I recommend doing this only if using a segment model in your current environment does not produce any noticeable changes in the generated images.
If you have any questions about the forked adetailer, feel free to leave a comment and I will help as much as I can.
@newtextdoc11996 thank you.
I'm genuinely impressed with this model, it gives my pics really good details, tho something I liked was using the pussy separately so I can add pussy and clitoris for a really good looking pussy, either ways I like the results with the 5.0 so far, also I see it gives nipples to testicles but ig it can be corrected or sum, if this keeps improving boy it'll be really good! Very nice work OC!!!❤️
Thank you for the positive feedback! I really appreciate it.
About the part "also I see it gives nipples to testicles,"
if the appearance changes after the Detailer, please try adjusting the parameters using the FAQ section titled "How to Deal with the Issue of Detected Areas Being Replaced by Another Object."
最近になって使い始めましたが、検出がとても良くできます。ありがとうございます。
できれば亀頭の部分の検出がよくなったら最高です。検討おねがいします。
（パイズリ中など亀頭だけ見えているときに検出率が低めでした）
フィードバックありがとうございます。
パイズリの画像はあまり含めていなかったので、次回アップデートの改善項目に追加しておきます。
要是有confyui的工作流就好了(◞‸◟ )
Please take a look at the workflow below for reference. It is almost the same as the one that I normally use.
https://civitai.com/models/1341828/sdxl-modular-workflow
https://civitai.com/articles/17339 you can look this
This model has a drawback, as its detection performance for dark pussy is very poor. Hopping this problem can be solved.
Thank you for the feedback.
Since it was not included in the existing dataset, I will add to next update.
Been trying to get this to work on Anima images and is oddly not detecting anything regardless of threshold settings.
It may not detect properly depending on the art style.
This model was mainly trained on images generated with WAI and Hassaku, so if the style is very different, detection might not work well.
Base models like Anima, where the art style is not very consistent, especially tend to have lower detection rates.
Hello, is it possible to have an example prompt for adetailer in automatic1111?
I want to sort prompt by [class=pussy][class=nipple], but it does not work? can you show me the proper format, please?
Can we have individual models please?
Also V5 nipples (aio) mistake lips for nipples more instead of nipples themselves.