Hassaku-based. The style itself isn't hard to achieve with prompts, but the art style looks nice, so I want to turn it into a LoRA for convenience. Only 35 training images because it doesn't need much variety.
I recommend using image enhancing and not highres, the images should be smoother that way.
