SDXL LoRA Inspired by the original <a target="_blank" rel="ugc" href="https://civarchive.com/models/52496?modelVersionId=56909">Hotsquat</a> by @WaifuWizardWongv0.2 retrained with Prodigy - MUCH better results.<a target="_blank" rel="ugc" href="https://tensor.art/models/625257185988909718?source_id=nzyzpFnql0G6pfUoYHr29xYv">&gt;&gt; Try it before you download &lt;&lt;</a><hr /><h3 id="heading-261">Training Data</h3><ul><li>62 images upscaled from SD1.5 generated images, 832x1216</li><li>1984 steps, 12 epoch, Prodigy</li></ul><hr /><h3 id="heading-84">Suggested Params</h3><ul><li>LoRA weight: 0.6-0.8</li><li>DPM fast or DPM++ 2M SDE Karras</li><li>7-9 CFG scale</li><li>Res: 832x1216</li></ul><hr />USAGE AGREEMENTThis took an incredible amount of time and effort; please give proper credit when using. Other than that, go wild!Donations always appreciated.<pre><code>ETH / Any ERC20: 0x00eFf0937fC7468f76C299fc4A3c22CC9432757A</code></pre>Leave a comment for a commissioned model.

Hotsquat XL [BETA v0.2]

<ul><li>Standard LoRA trained w/ AdamW8bit</li><li>Needs heavy negative prompting</li></ul>

Files with SHA256: 9a3278f5f66c0051d1565d0d1c20a12a98e0d049ab016d448ee54ddd738d0808