<h1 id="luckymix">LuckyMix🍀</h1> Motivation: Creating a model with expanded knowledge and a mostly neutral style that would be great for using with @Artists and style Loras. <ul><li>Recommended generation settings:</li><li>CFG: 5.0</li><li>Steps: 30</li><li>Sampler: er_sde</li><li>Scheduler: simple/normal</li></ul> <pre><code>Positive prompt: masterpiece, best quality, absurdres, highres</code></pre><pre><code>Negative prompt: score_1, score_2, score_3, worst quality, low quality, artist name, blurry</code></pre> V2: Improved style coherence. Leans into 2D with less shiny skin compared to previous version. V1: Responsive to e621 tags. Tends to add a bit of shadows and highlights. Might slightly lean into 2.5d. Enjoy!

LuckyMix🍀

Responsive to e621 tags. Tends to add a bit of shadows and highlights.Special thanks to Anima aesteticV1-1 aniMurr, WAI-ANIMA, and Nova Anime AM models.

Files with SHA256: c5e5f9b8f23995d0bd745b65c5e8f7530c62ddbfafb63ccc6470dc01f9cc5a4d