<h1 id="arbuzbudesh-style-lora">ArbuzBudesh Style LoRA</h1>Support ArbuzBudesh here:<ul><li><a target="_blank" rel="ugc" href="https://www.furaffinity.net/user/arbuzbudesh">https://www.furaffinity.net/user/arbuzbudesh</a></li><li><a target="_blank" rel="ugc" href="https://twitter.com/ArbuzBudesh">https://twitter.com/ArbuzBudesh</a></li><li><a target="_blank" rel="ugc" href="https://cenaf.newgrounds.com">https://cenaf.newgrounds.com</a></li><li><a target="_blank" rel="ugc" href="https://www.hentai-foundry.com/user/ArbuzBudesh/profile">https://www.hentai-foundry.com/user/ArbuzBudesh/profile</a></li><li><a target="_blank" rel="ugc" href="https://www.patreon.com/arbuzbudesh">https://www.patreon.com/arbuzbudesh</a></li></ul>Read The Price Of Freedom here:<a target="_blank" rel="ugc" href="https://priceforfreedom.net/comic/price_for_freedom/1">https://priceforfreedom.net/comic/price_for_freedom/1</a>Get the game here:<a target="_blank" rel="ugc" href="https://priceforfreedom.net/">https://priceforfreedom.net/</a><h2 id="disclaimer-and-request">Disclaimer and Request</h2>Always credit ArbuzBudesh when posting images generated with this LoRA.Trained on 100% publicly available images taken from e621.Please do not sell this LoRA, or images generated with this LoRA (additionally, do not sell access to this LoRA, or access to images generated with this LoRA, e.g. via Patreon or other paid/subscription services).Please do not remove metadata from images generated with this LoRA.<h2 id="model-info">Model Info</h2><ul><li>Trained on: <a target="_blank" rel="ugc" href="https://civarchive.com/models/43784?modelVersionId=48423">CoffeeNSFW</a> (My fav. anime model)</li><li>Epoch: 10</li><li>Repeats: 5</li><li>Resolution: 640</li><li>Images: <a target="_blank" rel="ugc" href="https://e621.net/posts?tags=arbuzbudesh+-flash+-video+-animated+-filetype%3Agif+date%3A%3E01%2F01%2F2016">470 from e621</a><ul><li>I specifically filtered out images with the following tags:<ul><li>video</li><li>animated</li><li>filetype:gif</li><li>date: older than 01/01/2016</li><li>lowres</li></ul></li></ul></li><li>Steps: 23500</li><li>ClipSkip: 2</li></ul><h2 id="recommendations">Recommendations</h2><ul><li>Trigger Words: None needed</li><li>Weight: Around 0.9 yielded good results</li><li>Samplers:<ul><li>Euler A showed surprisingly strong</li><li>Otherwise DPM++ 2M Karras was my go-to</li></ul></li><li>Base Resolutions: Around 384x576</li><li>Upscalers:<ul><li><a target="_blank" rel="ugc" href="https://de-next.owncube.com/index.php/s/x99pKzS7TNaErrC/download">4x_fatal_Anime_500000_G</a> showed the best results</li><li><a target="_blank" rel="ugc" href="https://upscale.wiki/wiki/Model_Database#Universal_Models">4x_fooldhardy_Remacri</a> also did OK, but more washed out</li><li>Latent (nearest-exact) is also an OK option, but only at very high denoise</li></ul></li><li>Extensions:<ul><li>Use <a target="_blank" rel="ugc" href="https://github.com/mcmonkeyprojects/sd-dynamic-thresholding">Dynamic Thresholding Extension</a> to use high CFGs without burn-in</li></ul></li></ul>

Files with SHA256: 76cd4202845fb51b358f9601c94fa48d292c60432c41c0fb1733adb1609a6ff4