This lora is inspired by Korean novels artwork, hence the name. It is strictly trained from my other lora Porcelana but with a custom prompt, I just wanted to set aside that specific prompt to create a new set of lora to mix and match. This is why this lora is way too glossy or just have great lighting like Porcelana has. Basically it's very cinematic, I can't say it's a better version than those Rofans but I did mimic the style!

Recommended sampling steps by testing: 12-14. But you can also do 8-10 if you don't have longer prompts because the more the longer your prompt is, the more chances the brush strokes would be slightly blurred. Like yes it has high resolution but it'll act like it doesn't lol.

I realized it lowkey prefers shorter prompts? Idk. Currently there's a bug with couple portraits which is why I'm still retraining it. However solo portraits seem fine especially women!