Title: illustraVERSA – (Illustrious Merge)
Description:
illustraVERSA is a custom merged checkpoint combining Arthemy Comics V4.5 (Illustrious) and Animij v6 to create a versatile, high-quality illustration model. This merge preserves the broad prompting understanding and stylistic flexibility of Arthemy’s checkpoint while integrating Reijlitas’s refined details, offering exceptional results across a wide range of CFG values (from 8 to even above 10).
Whether you’re creating comic-style illustrations, bold linework, or expressive character art, illustraVERSA delivers consistent quality and detail, making it a reliable choice for both experimental and polished outputs.
Credits:
@Arthemy
https://civarchive.com/models/1273254?modelVersionId=1963116@reijlita
https://civarchive.com/models/1353314?modelVersionId=2119925
This model contains a different VAE
https://civarchive.com/models/1684460/illustrious-xl-v20
On version 2.0
Credit to
@phinjo
https://civarchive.com/models/879396/commix-pony-illustrious
Notes:
This model is intended for use in ComfyUI with SDXL-based workflows.
LoRAs used are KFC secret recipe
I will gift buzz if any one nails it on their images
:D
This is my first merge, I will be testing and fixing things to bring a V2 soon.
This one is a hail mary for now
illustraVERSA-AE model
This version is more Anime than comic
Improved faces they are more feminine
removed texture (crosshatching)
this one implements
@Arthemy – https://civarchive.com/models/1273254?modelVersionId=1963116
and
@NatMontero - https://civarchive.com/models/1859774/natty-illustrious
Description
Toned down the linework. Smoother style not to overly saturated/overcooked with outlines.
FAQ
Comments (3)
Another Banger!!! Works really well with the DMD2 setup too!
For anyone experimenting with that workflow, try a Euler Exponential variant with 10 steps Cfg 1
(Euler SMEA DY - Exponential is producing some nice results)
Welp. Even with this new version, I haven't been able to replicate these.
(https://civitai.com/images/96637916 / https://civitai.com/images/96579051)
Either way, this is a fantastic checkpoint if you don't plan on using any LoRas. I'm not getting the best results by mixing LoRas, however, this is AMAZING as a standalone (applying only embeddings and intricate text prompts). Good stuff!
I made those img 2 img with controlNET hehe a few weeks ago I updated my workflow to txt 2 img