Built for lighting and gravity. Deep shadow, bold highlight, and characters that carry real weight in their expression; this LoRA is a study in chiaroscuro, the classical technique of strong light-and-dark contrast to create drama and depth. It's not locked to any one mood. Push it shadowed and heavy, or push it warm and vivid - the dramatic lighting holds whichever way you direct it. Trigger word: <code>darkmeridian</code> - place it first, before everything else. I get best results using minimal Illustrious based quality tags (ex: masterpiece, absurdres and that's it)Recommended checkpoint: <a target="_blank" rel="ugc" href="https://civarchive.com/models/1034922/dasiwa-illustrious-or-anime">Dasiwa Epitaph Ecstasy</a> No negative prompt = Painterly/2.5d Negative Prompt = clean cel-shading Recommended negative prompt: (remove photorealistic/realistic for a more 2.5d look) <code>worst quality, low quality, blurry, extra fingers, extra digits, malformed hands, bad anatomy, distorted face, deformed face, 3d, realistic, photorealistic, render, cgi, glossy, modern, contemporary</code> <img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6726029e-0744-4937-a523-4cd88352f55c/original=true/6726029e-0744-4937-a523-4cd88352f55c.jpeg" /><hr />Suggested Settings:Euler A / Karras or Beta57 CFG 5.0 28–30 steps Recommended LoRA weight - .90 - 1.0<hr />Negative prompt: <code>worst quality, low quality, blurry, extra fingers, extra digits, malformed hands, bad anatomy, distorted face, deformed face, 3d, realistic, photorealistic, render, cgi, glossy, modern, contemporary</code><hr />All gallery images generated straight from ComfyUI. No post processing, upscaling or inpainting (except where noted).

Files with SHA256: 506891690ae78c203063fb639a1186001fbace0d5283ebf01bcbc90fb3ca93c2