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
その他

Giving/Getting Wedgies

#食い込み
AIツール
  • AIアニメ生成
  • アニメフィギュア生成
  • キャラクターシート生成
  • 雑誌表紙生成
  • ぬいぐるみ生成
  • 巨大彫像生成

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
Giving/Getting Wedgies - AI Modelのカバー画像
作成者
Avatar image of Daiz
avatar-frame

Daiz

自作LoRAを学習
アップロード日時
2026/01/21 4:40
採用数
26.1k
獲得
490,311
レビュー
普通  (9)
トリガーワード

wedgtim, two_people, person_giving_wedgie_to_another, giving_wedgie, receiving_wedgie, underwear_pull

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

Hyperpregnancy XL - AI Modelのカバー画像

ユーザーLoRA

Hyperpregnancy XL
368
Usagi Tsukino 90s Style V2 - AI Modelのカバー画像

ユーザーLoRA

Usagi Tsukino 90s Style V2
5
Mavuika - AI Modelのカバー画像

ユーザーLoRA

Mavuika
42
Isolde Zenless Zone Zero/ ZZZ - AI Modelのカバー画像

ユーザーLoRA

Isolde Zenless Zone Zero/ ZZZ
3

詳細説明

Made this to make it easier to generate images with a character giving another character a wedgie. Though just like the "Group Hanging Wedgies" Lora it can be hard to use and often require to have to put in some work or do multiple attempts to get it right. Start around 1.0 weight and then heighten or lower depending on Lora combinations, higher weight usually resulting in more drastic wedgies but can also make the result drop in quality. As always, putting multiple character Lora's together often results in design bleeding. I've found a few methods that can help alleviate this somewhat, not using character Loras and just the character names usually prevents it and if the character is popular enough is often all you need. If you have to use Character Loras though, I've found a trick that can help: first use the character loras like normal, once you get a decent result/pose you can make that result into the reference image, then drop the character loras (but make sure to keep the names and clothing prompts) and generate with that, this will often start to fix the bleeding provided they weren't too drastic. This trick can sometimes also work in reverse, staring with no Loras then adding the Character Loras after you have a base image ready. Another tip is to use proper separation, keeping each character separated by things like parenthesis, brackets, or the word BREAK. You can check the prompts of my multi-character posts for some inspiration, I've tried multiple ways but haven't fully decided on which is the best. If anyone has any methods they know, feel free to share in the comments. This is also the second attempt at this Lora (hence why it says 2.0 when you use it), the "beta" version is still there in the versions list if you want to try it, it should be worse overall but I did get the rare banger once in a while when testing it so maybe it can be of use. Also, one last thing. This Lora can technically work with just a single character as well. Just add "Solo" so it won't make a second character. You can also put "Solo Focus" and it can have character being wedgied by someone off-screen.

コメント