<h1 id="heading-7264">Kohaku XL alpha</h1><h3 id="heading-7265">An anime SDXL model trained on 100kπ images.</h3><p></p><hr /><h2 id="heading-7266">Introduction</h2><p>This model is trained from sdxl-1.0-base and merged with some other anime models to get optimal result.<br />I'm planing more training on more images (around 1.6M imgs) and some style lycoris models. so now it is just alpha version.<br /></p><hr /><h2 id="heading-7267">Usage Details</h2><p>This model is very flexible on resolution, you can use the resolution you used in sd1.x/2.x to get normal result(like 512x768), you can also use the resolution that is more native for sdxl(like 896*1280) or even bigger (1024x1536 also ok for t2i).</p><p></p><p>recommended negative prompt for anime style:<br /><code>photorealistic, 3d model, cinematic, bad anatomy, blurry, disembodied limb</code></p><p></p><hr /><h2 id="heading-7268">Acknowledgements</h2><p>Datasets</p><ul><li><p>280k imgs from <a target="_blank" rel="ugc" href="https://huggingface.co/MyneFactory">Mynefactory</a></p></li><li><p>30k imgs from <a target="_blank" rel="ugc" href="https://civarchive.com/user/alea31415">cybermeow</a></p></li></ul><p>Models</p><ul><li><p><a target="_blank" rel="ugc" href="https://civarchive.com/models/129329/aio-fantasticanimexxxl">AIO-FantasticAnimeXXXL</a></p></li><li><p><a target="_blank" rel="ugc" href="https://civarchive.com/models/118406/counterfeitxl">CounterfeitXL</a></p></li><li><p><a target="_blank" rel="ugc" href="https://civarchive.com/models/136719/nekorayxl">NekoRayXL</a></p></li></ul>

Files with SHA256: 82bc6f6ae9a469e2b541e4b784c670fc3dde4b587d2e690379754dc38bd1f14b