An attempt at making a "style"-type LoRA. Definitely not a 1-to-1 copy of the artist's style, but the result is decent enough, I think.Trigger word: magmadoughTested at weight 0.8, but other values may work as well.<h3 id="about-the-versions">About the versions</h3><ul><li>V1 - initial training</li><li>V2a - much less artifacts, but features a less pronounced style. Best used with any other LoRAs.</li><li>V2b - may artifact, but is closer to V1 visually. Behaves a bit like an overtrained model, so you might not get the best result on your first try.</li></ul>

MagmaDough style - Illustrious

Trained on AI-Toolkit, closer to the style of V1.

Files with SHA256: c2bab1e1f762d863eb0ef56a723e826635b463a94b4953ef4d0af30ffa0c33ed