weight 1 is best. (Base model is Noob-Illustrious SDXL)<h3 id="this-lora-only-works-with-the-illustrious-xl-series-checkpoints.-2pgdtgl8x">This LoRA only works with the Illustrious XL series checkpoints.</h3>main prompt <code>2girls, multiple girls, 1boy, ffm threesome, uncensored, penis, pubic hair, fellatio, deepthroat, (face in ass:1.2), anilingus, from side</code>sub prompt <code>x-ray</code> Possible prompt swap <code>upper body</code> ↔ <code>kneeling</code> ↔ <code>squatting</code><ul><li>Most images in this model dataset are landscape resolution. When using this model, it is recommended to create images in landscape resolution(or square resolution). It is not recommended to create images in portrait resolution.</li><li>このモデルのデータセット画像の多くは横解像度です。 画像を生成するときは、横解像度(または正方形の解像度)の画像で生成することをおすすめします。 縦は成功率が低いです。</li><li>학습 이미지의 대부분이 가로 해상도였기 때문에 이미지 생성할 땐 가로 위주(또는 정사각형 사이즈)의 이미지로 생성하는 것을 추천함. 세로도 되긴 되는데 성공률이 낮음.</li></ul>In the sample images, used hires.fix and the extension Adetailer.

Files with SHA256: bcbd736bdcf4e67de625bc9477341604f52e15d10adcef4a3e2473edc6ef55e8