Tutu's erotic uniform (traditional maid outfit)/图图的情趣制服（传统女仆装）

Created with Tutu AI Workflow

Tagged with Tutu Annotator. Trained with Tutu Trainer.

This LoRA is now organized under my Tutu AI workflow. Tutu Trainer is used for LoRA training, model resource management, parameter presets, and training checks. Tutu Annotator is used for dataset annotation, prompt reverse engineering, English-friendly tags, and cleaner searchable captions.

For clothing LoRAs, clear outfit tags help users reproduce the exact costume instead of getting only a vague maid style. This model needs the apron, sleeves, headdress, shorts, thighhighs, thigh strap, and garter straps to work together.

• Official website: https://zhaotutu.xyz/
• Tutu Trainer: https://zhaotutu.xyz/downloads/tututrainer/
• Tutu Annotator: https://zhaotutu.xyz/downloads/tutuannotator/
• Bilibili: https://space.bilibili.com/431046154
• YouTube: https://www.youtube.com/@zhaotutu/videos
• Telegram: https://t.me/zhaotutu
• TikTok: https://www.tiktok.com/@zhaotututu
• Reddit: https://www.reddit.com/user/zhaotututu/
• GitHub: https://github.com/zhaotututu
• Email: [email protected]
• QQ: 331506796

Tutu Uniform: Traditional Maid Outfit

This LoRA focuses on a more traditional and authentic maid outfit. Unlike many of my difference-extraction experiments, this one was made as a more conventional training test using collected reference images.

I wanted to see whether the normal training route could still produce a solid clothing model. In practice, the result was good, with the LOSS value around 0.03.

• Type: traditional maid outfit / clothing LoRA
• Base model: SD 1.5
• Trigger core: tutututu
• Recommended LoRA weight: 0.4-0.8
• Do not exceed: about 0.85 unless the base model needs it
• Recommended CLIP skip: 2
• CFG Scale: start around 7

Full trigger prompt

For better restoration, use the full outfit group:

tutututu, apron, long sleeves, maid, maid apron, puffy sleeves, maid headdress, black shorts, black thighhighs, thigh strap, garter straps

What this model is good at

• Creating a traditional maid outfit rather than only a generic maid impression.
• Keeping apron, long sleeves, puffy sleeves, and maid headdress together.
• Adding black shorts, black thighhighs, thigh strap, and garter straps when prompted.
• Working well across many SD 1.5 base models after weight tuning.
• Providing a more normal training-process example inside the Tutu uniform series.

Recommended settings

• Make sure the LoRA is loaded correctly.
• Use the full trigger prompt when you need better outfit restoration.
• Recommended CLIP skip / CLIP stop layers: 2.
• LoRA weight usually works between 0.4 and 0.8.
• It is recommended not to exceed 0.85.
• Start CFG Scale around 7.
• Do not add too many negative prompts, or the LoRA effect may become weaker.
• If a mosaic or broken texture appears after increasing weight, lower this LoRA weight first, then lower other active LoRAs if needed.

Prompt keywords

• traditional maid outfit
• maid
• maid apron
• apron
• long sleeves
• puffy sleeves
• maid headdress
• black shorts
• black thighhighs
• thigh strap
• garter straps

FAQ

Why does it look like a generic maid outfit?

Use the full trigger prompt and strengthen maid apron, puffy sleeves, and maid headdress.

Why do the lower-body details disappear?

Add black shorts, black thighhighs, thigh strap, and garter straps.

Why does the image break after increasing weight?

Keep the weight under 0.85 first. If multiple LoRAs are active, reduce the total LoRA pressure.

Responsible use

• You are fully responsible for any creative work made with this model.
• Do not use this model to create or share illegal, harmful, malicious, defamatory, fraudulent, or political content.
• If you use this model commercially or need custom LoRA cooperation, please contact me first. Thank you.

中文说明

这个 LoRA 的主题是比较传统、比较正宗的女仆装。和我很多差异提取实验不同，这个模型是一次更常规的训练测试，使用收集到的参考图来制作。

我当时就是想试试：正常制作模型、正常训练服装，质量到底能不能做好。实践证明效果还不错，LOSS 值在 0.03 左右。

• 类型：传统女仆装 / 服装 LoRA
• 基础模型：SD 1.5
• 核心触发词：tutututu
• 推荐 LoRA 权重：0.4-0.8
• 建议不要超过：0.85 左右，除非底模确实需要
• CLIP 终止层：2
• CFG Scale：从 7 左右开始

完整触发词

为了更好还原服装，建议使用完整触发词组：

tutututu, apron, long sleeves, maid, maid apron, puffy sleeves, maid headdress, black shorts, black thighhighs, thigh strap, garter straps

这个模型适合什么

• 生成传统女仆装，而不是只有模糊的女仆印象。
• 让围裙、长袖、泡泡袖、女仆头饰保持在同一套服装里。
• 在提示词足够明确时，生成黑色短裤、黑色大腿袜、大腿绑带和吊袜带。
• 在多种 SD 1.5 底模下，通过权重微调取得不错效果。
• 作为图图制服系列里更常规训练流程的一个样本。

推荐参数

• 确保正确加载 LoRA。
• 如果需要更高服装还原度，请使用完整触发词组。
• CLIP 终止层数建议设置为 2。
• LoRA 权重通常在 0.4 到 0.8 之间都可以。
• 建议不要超过 0.85。
• CFG Scale 可以从 7 左右开始。
• 负面提示词不要堆太多，否则可能影响 LoRA 效果。
• 如果权重调高后出现马赛克或纹理破碎，先降低本 LoRA 权重，再考虑降低其他 LoRA 权重。

出图关键词

• traditional maid outfit
• maid
• maid apron
• apron
• long sleeves
• puffy sleeves
• maid headdress
• black shorts
• black thighhighs
• thigh strap
• garter straps

常见问题

为什么只是普通女仆感，不够还原？

请使用完整触发词，并加强 maid apron、puffy sleeves、maid headdress。

为什么下半身细节不见了？

请补充 black shorts、black thighhighs、thigh strap、garter straps。

为什么权重一高画面就坏？

先把权重控制在 0.85 以下。如果同时用了多个 LoRA，请降低整体 LoRA 压力。

声明

• 你需要对使用本模型生成的任何作品承担全部责任。
• 不要使用本模型制作或传播非法、有害、恶意、诽谤、诈骗或政治用途内容。
• 如果将本模型用于商业用途，或需要定制 LoRA 合作，请提前联系我，谢谢。

日本語説明

これは traditional maid outfit をテーマにした LoRA です。私の多くの difference-extraction experiments とは違い、このモデルは collected reference images を使った、より普通の training test として作りました。

普通の training route でも clothing model の品質を出せるか試したかったモデルです。結果は悪くなく、LOSS value は 0.03 前後でした。

• Type: traditional maid outfit / clothing LoRA
• Base model: SD 1.5
• Trigger core: tutututu
• Recommended LoRA weight: 0.4-0.8
• Do not exceed: about 0.85 unless needed
• CLIP skip: 2
• CFG Scale: start around 7

完全な trigger prompt

outfit restoration を良くしたい場合は、この full outfit group を使ってください:

tutututu, apron, long sleeves, maid, maid apron, puffy sleeves, maid headdress, black shorts, black thighhighs, thigh strap, garter straps

得意なこと

• generic maid impression ではなく traditional maid outfit を出すこと。
• apron、long sleeves、puffy sleeves、maid headdress を同じ outfit として保つこと。
• prompt に応じて black shorts、black thighhighs、thigh strap、garter straps を出すこと。
• SD 1.5 base model で weight 調整しながら使えること。
• Tutu uniform series の中で normal training process の例として使えること。

推奨設定

• LoRA が正しく読み込まれていることを確認してください。
• 高い outfit restoration が必要な場合は full trigger prompt を使ってください。
• CLIP skip: 2 を推奨します。
• LoRA weight は通常 0.4 から 0.8 で使えます。
• 0.85 を超えないことをおすすめします。
• CFG Scale: 7 前後から試してください。
• negative prompt を増やしすぎると LoRA 効果が弱くなることがあります。
• weight を上げて texture が崩れる場合は、まずこの LoRA weight を下げてください。

プロンプト例

• traditional maid outfit
• maid
• maid apron
• apron
• long sleeves
• puffy sleeves
• maid headdress
• black shorts
• black thighhighs
• thigh strap
• garter straps

FAQ

なぜ generic maid outfit になりますか？

full trigger prompt を使い、maid apron, puffy sleeves, maid headdress を強めてください。

なぜ lower-body details が消えますか？

black shorts, black thighhighs, thigh strap, garter straps を追加してください。

weight を上げると画像が崩れる場合は？

まず 0.85 以下にしてください。複数 LoRA を使っている場合は、全体の LoRA pressure を下げてください。