<h3 id="heading-482">Note: For best results, use with SDXL Turbo, or better yet <a target="_blank" rel="ugc" href="https://civarchive.com/models/215418">TurboVisionXL</a>.</h3><h3 id="heading-483">Use with keywords: upskirt, then optionally 'vagina out' if you want that, then 'rear shot' or 'front shot' depending on whether you want them shot from the front or rear.</h3><h3 id="heading-483">Extra keywords like 'lifting up her skirt', 'skirt blowing up in the wind', 'bending over', 'crouching', 'squatting', 'legs up in the air' etc. should also work.</h3><p></p><p>Trained on thousands of carefully selected high definition pantyless upskirt images with full detailed captions I wrote systematically but manually.</p><p>It generalizes surprisingly well, using the body of the subject rather than any specific details from the training dataset.</p><p>It seems to have learned the actual <em>concept</em> of an upskirt, which is what I was aiming for.</p><p></p><p>I've been sitting on this one since SDXL came out because there were still a lot of bugs I wanted to fix, but for some reason SDXL Turbo fixes them 95% of the time, probably due to turbo's lack of weird anatomy bugs that misalign nudes by default.</p>

Upskirt Diffusion XL

<p>Trained on 1024x1024 images.</p><p>I have trained a bucketed one at higher resolution which should be better in theory, but it was significantly more buggy.</p>

Files with SHA256: 9b8a9764360d7fdfa79879eaa32fb6897f5647bbd1f4bfeebda711cebab5f641