<h2 id="check-out-the-original-artist's-twitter-and-other-social-media.-their-artworks-are-much-better-than-anything-this-lora-can-generate.">Check out the original artist's <a target="_blank" rel="ugc" href="https://x.com/Maoh_isGod">Twitter</a> and <a target="_blank" rel="ugc" href="https://www.pixiv.net/en/users/4331804">other social media</a>. Their artworks are much better than anything this LoRA can generate.</h2><h2 id="you-can-commission-them-here.-(if-you-know-japanese)">You can commission them <a rel="ugc" href="https://skeb.jp/@Maoh_isGod">here</a>. (If you know Japanese)</h2>The dataset has been cleared out of any scat/fart/watersports related images. If you're specifically looking for those kind of things in this LoRA, I'm sorry to disappoint. I was only interested in the specific way the artist draws masked men, which I find appealing.I recommend using a weight of 8.0 or lower. I'm still not sure how to determine if a LoRA is overtrained or not. If any of you have more know-how about training than I do and you think this is overtrained, please let me know.It's not necessary, but the LoRA has a trigger word - "d-k-c". There might be some instances where the character's face, eyes and hands will render poorly. Adetailer is highly recommended as always.

Files with SHA256: 10708464e3131bdfbcfd4b458031b4ac4077c088296d12affeffc78f8577a9e5