<h1 id="model-description">Model Description</h1><h3 id="0.1-preview">0.1 - Preview</h3>Attempt to sloppify Mugen SDXL, was trained with 3.7k images actually handpicked (unlike other liars) for 150 epochs on 1536 target resolution.(this is a preview version, therefore it's allowed to be bad)All samples are RAW GENSthings to be aware of before complaining:<ul><li>default gens are kinda meh, pupils started to appear properly, but it has a long way</li><li>hands, 30% chance to fail (mainly finger count), depends on artist.</li><li>specific gens on 1536 resolution will get stretched due to not being present in the dataset, happens very rarely.</li><li>USE ARTISTS</li></ul>Usage + what I used in the sample gens:go to <a target="_blank" rel="ugc" href="https://civarchive.com/models/2237480/mugen-sdxl-with-flux2s-vae">Mugen</a> and make sure you can run it<ul><li>shift = 8.0</li><li>scheduler = sgm_uniform</li><li>sampler = euler_a2/euler_a/dpmpp_3m_sde_gpu/heun</li><li>steps = 25</li><li>cfg = 8.0 (can go wild here)</li><li>resolutions (works both ways portrait/landscape): 1536x1536 (1:1), 1344x1792 (3:4), 1280x1856 (2:3), 1152x2048 (9:16), 1024x2368 (9:21) 1024 (works the same)</li><li>nice artists/themes that you can prompt = rella/horiguchi yukiko/murota yuuhei/catbell/anime coloring</li></ul><h2 id="0.3-preview">0.3 - Preview</h2>everything improved, more pupils, you tell me, use artists<h2 id="0.5-preview-(coming-soon)">0.5 - Preview (coming soon)</h2>added more artists to the mix, biasing eyes, keeping artist tokens to force model to override bad old art, added repeats for under-represented artists, going for double training amount

Files with SHA256: 926df40432ff67cdcb5b0739fb601fa96f63c6902483206dc3fa72b18c6608c0