<h3 id="this-model-is-not-great-but-i'm-currently-working-on-some-improvements-and-hopefully-another-version.-unfortunately-for-me-that-means-tagging-another-900+-images-so-it'll-be-a-while.-my-hope-is-that-i-can-reduce-artifacts-and-improve-image-quality-significantly.-x6i2wui7z">This model is not great, but I'm currently working on some improvements and hopefully another version. Unfortunately for me, that means tagging another 900+ images, so it'll be a while. My hope is that I can reduce artifacts and improve image quality significantly. </h3>This LoRA is an amalgamation of over 900+ hand-tagged and curated images, including photographs and illustrations. Rather than teaching one concept, I wanted one LoRA that encompasses most of my tastes in porn.It has been heavily trained (20 epochs on 900+ images) on all of the following categories:<ul><li>Symbiotes</li><li>Transformation</li><li>Corruption</li><li>Demon girls</li><li>Futa</li><li>Femboys</li><li>Glowing genetalia</li><li>Long tongues</li><li>Colored tongues</li><li>Latex</li><li>Gradient skin (we all know it's just peak character design)</li><li>penis bulges/covered penises</li></ul>As many of these are somewhat niche, many of the images used to train were low quality. For all images that didn't get an A+ in the art department, but got an A++ in the kink department, I assigned the tag 'lowquality'. Be sure to put this in your negative prompt, it will likely significantly enhance the quality of your art.It's 3:00 AM and I have work in the morning, I'll add some extra detail and pictures in the damn morning.

Files with SHA256: 3fa7f51dc5a5b78b104743cbb8ad399aa18bb85eebc55227fe94fbc50c90bfbd