Saw the bounty at https://civarchive.com/bounties/4590/bai-asuka-art-style
As with all art style LoRA/LoCons, if you are the art style owner and want to take this down, do let me or the CivitAI team know. Also note that commercial use is not allowed for this LoCon - for proprietary reasons, so please do not sell AI generated mangas based on their art style.
Let me know what worked well for you by posting your generations in the gallery.
Description
Trained on Pony Diffusion V6 XL, with a cleaner dataset and LoCon parameters.
There are no trigger words used in this version, but following the prompting "style" in the gallery showcases will give the best results. In addition:
Use a LoRA weight scale between 0.9 to 1.0
The higher the closer it is to Bai Asuka's style, but will clutter the image with speech bubbles, sfx, etc. This is because I wasn't able to clean the dataset much. (my bad)
Use detailers such as EasyNegative