<a target="_blank" rel="ugc" href="https://www.pixiv.net/en/users/15039870">Juju</a> <a target="_blank" rel="ugc" href="https://twitter.com/jujunaught">(Jujunaught)</a> Style Lora.<hr /><h2 id="update:">Update:</h2>Noob version trained on epred 1.0. Can be wonky. Works well with Illustrious mixes too. Play around with the strength if results are questionable.<hr />This style can produce a lot of NSFW content. Which begs the question, should I post more images that are leaning heavily on that side of things (full blown sexual acts, fetishes, porn physics, etc.)? Or should I keep things mild (nudity and genitalia only)? I'll post some of both in the first upload image examples, but judging by the likes/reactions and downloads I've been getting recently, people prefer hardcore.

Jujunaught Style [IL] | [noobE] | [PonyXL]

Trained on Illustrious v01. Use "pkjujunaught" as a trigger word for the images in the dataset and you can add "jujunaught" as well to use the latent knowledge of the checkpoint. That will strengthen the result but may introduce some wackiness and bias. Sidenote: It took me seven business and 3 weekend days to upload the model. CivitAI has the best server and site stability this side of the Milky Way. You LOVE to see it

Files with SHA256: f8934fc05c2fe8bb0c75368e442fb2c8327e2323f9df2a1921e4c478a925fa80