Update v5 (Final)
For everyone who downloaded the version from April 26 — you may want to re-download the updated version, but feel free to do whatever you prefer.
This is the final version — no further updates will follow.
I’ve pushed the quality and style as far as possible, delivering the most consistent and refined results I could achieve.
This LoRA represents the finished version of my work — enjoy creating with it.
No trigger word is required.
Recommended strength: 0.7 – 1.0
Hi-Res Fix and ADetailer are recommended.
P.S.: Images were generated using Liquid111VAE.
Description
Update v2.5
This version was trained again on WAI-Illustrious-SDXL v16, but with additional training material.
I added 40+ more images to the dataset to improve consistency and overall quality