<h1 id="imminent-penetration-(and-after-sex)-8z8mag74n">Imminent Penetration (and after sex)</h1>Lycoris All sample images used it at .7IMPORTANT INFO : PLEASE READ<h3 id="this-lora-was-trained-on-gurilamash-xxx-sdxl-sqys1qpfc">This LORA was trained on <a rel="ugc" href="https://civarchive.com/models/533591/gurilamash-xxx-sdxl">GurilaMash XXX SDXL </a></h3><h3 id="it-will-only-work-with-my-model-or-another-bigasp-based-model-which-so-far-includes:-rgko9jv17">It will only work with my model, or another bigASP 🐍 based model, which so far includes:</h3><h2 id="bigasp-hob375vni"><a rel="ugc" href="https://civarchive.com/models/502468/bigasp-v1">bigASP</a>🐍<a rel="ugc" href="https://civarchive.com/models/502468/bigasp-v1"> </a></h2>and<h2 id="lustify!-60o10w43q"><a rel="ugc" href="https://civarchive.com/models/573152/lustify-sdxl-nsfw-and-sfw-checkpoint">LUSTIFY! </a></h2>This LORA WILL NOT WORK WITHOUT One of the above modelsdon't bother trying it on a pony model or any other sdxl model, I tested, it don't work. It only works, and it works VERY well, on bigASP and it's deriviatives.I have included the caption dataset as a separate download to provide a guide for prompting.Works best at .5-.8.9+ starts to reduce quality-.4 and it has little effect.It has watermarks. Deal with it. This lora was a proof of concept to see if I could train bigASP loras. I think you'll agree it was a success. Use "imminent penetration" high in the bigASP based tags prompt. Use "after vaginal" lower in the prompt if you want a lot of body fluids, it's too powerful, and often leads to penetration. The model sees vaginal as sex, and the "after" is an afterthought most of the time. It's not necessary to produce body fluids, they can be prompted by location as per bigASP prompts. Mine the captions, or copy-paste them, if you can't seem to get it right. This LYCORIS-LOCON was trained on Kohya_ss at 5-e4 linear for 45 epochs on a dataset of about 500 images. More is coming, very soon.

Imminent Penetration (and after sex) for bigASP🐍 based Checkpoints

500 images trained at 5e-4 linear 45epochs

Files with SHA256: 2f86011b4d0d4a439cb82846d984fcab3b17a605ca88587c11f115833c122403