Based on the artstyle of artist Coburaco (Cobura no Oyatsu).Tested on AutismMix and Prefect Illustrious XL.Preview images contain Comfy UI workflows and involve the use of Detailer nodes.IL 2.0 Update:Why so many variants?Because one model doesn't cover all cases, in my experience. Either the style influence is too weak or compatibility with other LORA is terrible. While testing, I found multiple epochs giving good results and figured there's no reason to discard perfectly good models, hence I'm sharing them all.Which one should I use?Balanced variant will work most of the times, but sometimes you might want this style in combination with 5 other LORA; that's when balanced kinda sucks; use soft or lite instead. Other times, you might want a simple image with a single LORA, in which case you can use hard or ultra to get a stronger resemblance than balanced. With larger sizes, you can get stronger style influence with better compatibility (XL balanced will have stronger style influence than M hard while also being more stable). Not hard and fast rules, FYI. Feel free to do your own testing.You can also check descriptions of the XL versions. The same description applies to L, M sizes and so on but with less style influence as they get smaller.If you're not sure, just get the balanced variant of whichever size you prefer 😉TLDR;<ul><li>Larger the size -&gt; stronger the influence</li><li>Lite, Soft -&gt; Works with maximum LORA</li><li>Hard, Ultra -&gt; Not as stable but stronger style influence</li><li>Extreme -&gt; Mostly unstable, would use with ControlNet only</li><li>Balanced -&gt; This is the one you want 90% of the time</li></ul>Enjoy 😊

Coburaco Style | Zilient

coburaco_md_hard.safetensors

v2.0 - M Hard

v2.0 - M Ultra

v2.0 - M Balanced

v2.0 - M Soft

v2.0 - M Lite

v2.0 - L Extreme

v2.0 - L Ultra

v2.0 - L Hard

v2.0 - L Balanced

v2.0 - L Soft

v2.0 - L Lite

v2.0 - XL Extreme

v2.0 - XL Ultra

v2.0 - XL Hard

v2.0 - XL Balanced

v2.0 - XL Soft

v2.0 - XL Lite

v1.0 (Illustrious)

v1.0

Files with SHA256: 998cacc5d52f6acaef5dc818fc01ff9e044bab8811a9b9aa4d3eba40b81610a3