Chromesthesia - CivArchive (CivitAI Archive)

(Capable of a wide range of styles, from flat to simi-realism).

Please Rate! - Your feedback, ratings or a follow are greatly appreciated. - Thankyou!

I used a face detailer in many of the example images, and most use upscaling (4x-nickelback). A face detailer is generally helpful depending on zoom level, If your subjects face is about 30% or more of the image it can be detrimental and cause loss of detail. I urge experimentation in all areas, adjust settings accordingly.

V2 - Baseline

I normally wouldn't publish a new version so quickly, but V1 felt a bit undercooked and sterile to me. I consider this a bit of a fix, It is a bit more toon biased, but prompting can still yield a wide range of styles. (See below for settings / v1 settings still apply).

V1 - Broadstroke

Suggested settings:

VAE: sdxl_vae (baked in)
Clip skip: 2 was used during merge, (setting 1 or 2 should yield same results)
Sampler: Euler Ancestral - SGM Uniform (most consistent good results). (more testing needed regarding additional samplers).
Resolution: all standard 1MP resolutions work well in portrait and landscape, (depending on context) (I often use 1024x1360 and 1120x1440 in portrait and landscape)
CFG: 2.8 - 9.0 (lower for more realism, (generally) - I commonly use 3.8, 5 ,7&9)
Steps: 22 - 36(I use 26 most often)

Prompting:

Primarily Danbooru tags, mixed with a bit of natural language prompting. (look to my sample images for examples and general formatting).

number of characters (1girl/1boy), character name, artist prompt, scene/environment/camera angle, action, expression, items, quality prompts (masterpiece, best quality)

negative prompts - lowres, worst quality, (it can be helpful to add a few things like "toon, illustration, unreal", if you're pushing for greater realism). (Additionally you can try the opposite I.E. adding photorealism to your neg prompt will give you a more flat / anime look).