This LoRA produces a couple dancing in formal wear. Illustrious is pretty good at keeping characters distinct, so you can usually just prompt for two characters and get them dancing, but don't expect perfection as this model was trained on one image. Seems to work best on hetero couples and also performs pretty well generating two women, but it seems to need more data to consistently to generate two men.This LoRA is associated with the article <a target="_blank" rel="ugc" href="https://civarchive.com/articles/3021/one-image-is-all-you-need">One Image Is All You Need</a>. It uses a single image dataset. Go read the article for more information on how to use these properly.<a target="_blank" rel="ugc" href="https://civarchive.com/user/ZaphodScotsman">@ZaphodScotsman</a> requested that I retrain <a target="_blank" rel="ugc" href="https://civarchive.com/models/786703/one-image-is-all-you-need-example-loras-pony-edition">https://civarchive.com/models/786703/one-image-is-all-you-need-example-loras-pony-edition</a> on Illustrious, and the results were better than I was expecting (again). So, here it is.Earlier I was planning on having this one continue to be part of the series of One Image Is All You Need example LoRAs, but this one has been wildly popular and I also have the SID collection now, so I'm going to have it stand alone from here on out.Trigger Word: <code>ballroom dancing</code>Helper Words: <code>from side, man, woman</code>Not trained but recommended: <code>1girl, 1boy</code>Full training caption: <code>ballroom dancing, from side, woman, pink dress, choker, dark brown hair bun, blush, back, man, white coat, short black hair, black tie, black pants, belt, theater stage background, bokeh</code>Unlike Pony, almost all of the training has been invested in the <code>ballroom dancing</code> caption. Using all the helper words <code>from side, man, woman</code> will generate happy couples embracing (occasionally sexually, but typically not) without the formal wear/dancing aspects.If using <code>ballroom dancing</code> it is recommended that you use this at lower strengths (&lt;=0.6) because otherwise a lot of the original image comes on through (unless that's what you want of course). Without that caption the overall effect can be weaker and so higher strengths work well.Training data included. Regularized with Oatd and Galena Bikini Pinups styles with <code>oatd</code> and <code>galena</code> style captions, but they're pretty weak.

Ballroom Dancing/Happy Couple (SID) Model (Illustrious)

Trained on Illustrious v1.0. 1024x1024 image as training data with 1024x1024 and 1536x1024 regularization data.

Files with SHA256: 1dc7d5a13454c43bc4d4e6a551a0d18f5b5258506245316d8852b56ea2ec32f9