<h3 id="chubby-cheeks-(now-with-a-different-dataset)-el28jmo8m">Chubby Cheeks (now with a different dataset)</h3><hr /><h1 id="v2.1-u1r33s40o">V2.1</h1>Note: More stable and using IL 0.1 as trained model. Also more authentic to v1.0, use this over v2.1.<h1 id="recommended-settings-j2qogyd0c">Recommended settings</h1>Strength: 1.0CFG: 3-8Steps: 20-30Sampler: dpm++ 2m sdeScheduler: karrasTrained on/best with: Illustrious XL v0.1 or any Finetune or Merge (Wai, Hassaku, Nova,)Positive:<pre><code>ODd5, 1girl, solo, eye focus, blush,</code></pre>Negative:<pre><code>artist name, twitter username</code></pre><hr /><h1 id="v2.0-snjkr92o0">V2.0</h1>Note: WAI v130 seems to have far more saturated colors and cooks the loras. I suggest using 3-5 CFG<h1 id="recommended-settings-cktoq14t5">Recommended settings</h1>Strength: 1.0CFG: 3-5Steps: 20-30Sampler: dpm++ 2m sdeScheduler: karrasTrained on/best with: WaiNSFWIllustrious V13.0Positive:<pre><code>1girl, solo, eye focus, blush, anime screencap, no lineart,</code></pre>Negative:<pre><code>artist name</code></pre><hr /><h1 id="v1.0-xt4arwq5o">V1.0</h1><h1 id="recommended-settings-ctuup5i0c">Recommended settings</h1>Strength: 0.6-1.0CFG: 5-7Steps: 20-30Sampler: dpm++ 2m sdeScheduler: karrasTrained on/best with: Illustrious XL v0.1 or any Finetune or Merge (Wai, Hassaku, Nova,)Positive:<pre><code>1girl, solo, eye focus, simple background, anime screencap, limited palette, no lineart, sketch</code></pre>Negative:<pre><code>Artist name</code></pre>

Files with SHA256: 4b4be7cb6974d8e6d25bdfb72a478a0a5ada137b962c96f6cd6fae2c8209dd1a