Purpose: The focus of this style mix is a painterly, soft shaded furry aesthetic that emphasizes gentle lighting, soft bodies and softer fur. Its only mildly painterly as I feel too much of that can lead to noisy looking images with illustrious. The main focus is on the softness of the shading and fur. The Blend: <a target="_blank" rel="ugc" href="https://civarchive.com/models/1178677/snowskau-style-illustrious">Snowskau Style [Illustrious]</a> <a target="_blank" rel="ugc" href="https://civarchive.com/models/514442/blotch-artstyle">Blotch_-_Art_Style</a> <a target="_blank" rel="ugc" href="https://civarchive.com/models/1581092/ekzonzz-artist-style-illustrious">Ekzonzz Artist Style - [Illustrious]</a> <a target="_blank" rel="ugc" href="https://civarchive.com/models/1227208/bubblewolf-style-noob">Bubblewolf style NOOB</a> <a target="_blank" rel="ugc" href="https://civarchive.com/models/2264925/honovy-style">Honovy Style</a> <a target="_blank" rel="ugc" href="https://civarchive.com/models/1076338/zaush-artstyle-ponyillustrious">Zaush_-_Art_Style [Pony/Illustrious]</a>Helpful Tags: soft shading– Pushes the rendering toward stronger soft-shading look. painterly – Pushes the rendering toward stronger painterly look. Although I find this tag can be quite weak. outline– Adds stronger outlines to characters, can help them stand out a little more and IMO make their fur look softer sometimes. Special Thanks and Extra Notes: A massive thanks to @Piper260, @Murf_ and @sxus_Sw for graciously letting me use their style LoRAs. Do check out the LoRAs listed above that are used in this mix, each of them are great and there's a reason why they were chosen.In my testing the <a target="_blank" rel="ugc" href="https://civarchive.com/models/1791766/indigo-latte-mix">Indigo Latte Mix</a> works best with this LoRA. What is an AnthroBlend Style: AnthroBlend styles are LoRA mixes made by yours truly, aimed at achieving a specific visual style without copying or mimicking any one artist directly. It’s not just a matter of stacking a bunch of similar style LoRAs together and hoping for the best. Each LoRA is individually tuned using block manipulation, which lets me push the parts of the style I want, like shading, line quality, or color behavior, while minimizing things like character design quirks or other stylistic noise baked into the original LoRA. Once the individual pieces are behaving the way I want, they’re combined and tested with carefully chosen weightings. After that, the whole stack is run through a <a target="_blank" rel="ugc" href="https://github.com/ethanfel/ComfyUI-LoRA-Optimizer">lora optimizer</a>, which adjusts how the LoRAs modify the underlying model weights so they work together instead of stepping on each other’s toes, preventing that muddy look that often happens with multiple LoRAs. The end result is a style that’s clean, consistent, and much easier to use than juggling a pile of separate LoRAs yourself.

Files with SHA256: 4d45f5a7da42672767f4d63e878816339970fd1079f784b013ab7b9ea1d1da81