PixAI会員
PixAI
PixAI
Generate
Sign in
ホーム
生成
モデル
Studio
NEW
コンテスト
Mio.2 で作成
ツールボックス
NEW
ガイド

ユーザーLoRA

XL
服装

Painted Clothes

AI Tools
  • AI Anime Generator
  • AI画像生成
  • AI動画生成
  • Anime Figure Generator
  • Character Sheet Generator
  • Magazine Cover Generator
  • Plushie Generator
  • Giant Statue Generator

探索

  • 人気タグ
  • ランキング
  • モデルマーケット
  • コンテスト
  • お知らせ

PixAIについて

  • ガイド
  • PixAI ブログ
  • Tsubaki.2
  • マスコットのミオに会う
  • コンテンツ規則

料金とヘルプ

  • 会員
  • クレジットパック
  • お問い合わせ

アプリ

  • PixAI アプリを入手
  • App Store
  • Google Play
© 2026 PixAI
  • 利用規約
  • プライバシーポリシー
  • 著作権ポリシー
  • クレジット表記
  • 特定商取引法
Painted Clothes - AI Model cover image
作成者
r

Ruby

自作LoRAを学習
アップロード日時
2025/11/07 16:47
採用数
1.0k
獲得
131,577
レビュー
不評
トリガーワード

skintight clothes, only wearing body paint, painted on clothes

権限
  • 画像生成と共有を許可する
  • ユーザーによるモデルのダウンロードを許可する
  • 生成画像の商用利用を許可する

おすすめLoRA

Redhead tavern wench - AI Model cover image

ユーザーLoRA

Redhead tavern wench
5
Riddle Rosehearts - AI Model cover image

ユーザーLoRA

Riddle Rosehearts
18
Asagiri Yoko - Genmu Senki Leda (SDXL)  - AI Model cover image

ユーザーLoRA

Asagiri Yoko - Genmu Senki Leda (SDXL)
12
quality me 2 - AI Model cover image

ユーザーLoRA

quality me 2
3

詳細説明

The goal here is to have painted on clothing that mimics real clothes. Works best with already tight clothing (hero spandex, yoga/work out clothes, swimsuits, etc). Was trained on predominantly single female character images so it has a strong bias for that. Adding NSFW to the negative prompt makes the clothes just skintight rather than painted.

For more visible features under the paint adding more specific secondary details seems to work best. Areola, clitoris, and labia have worked best in my experience, and adding them to negative prompts leads to a more featureless figure.

Occasionally there are random cutouts in the clothing, and adding to the negative prompt tags like cut outs, crotchless, clothing aside, panties aside, or any other tags related to clothing being present, but not covering everything, correct the issue.

Because of the scarcity of images to train this on the data set is predominantly made of images cherry picked from ones I generated. This means there is a very strong bias to the particular style I found that worked best, that being of the artist Naturalton (https://www.pixiv.net/en/users/232077). So if you like the aesthetic of the images that this generates then please check out his work, it is a million times better than anything this will make.

コメント