A traditional artist shared their dataset of images on Reddit, and I wanted to try training a style model on it.

No actual trigger word was used in training, so it's up to you to manipulate it with the prompt.

You may add color, or black-and-white to the prompt to try and steer the model.

This one is a lineart-like scribbly illustration drawing style.

I found the results when combined with Z-Image-Turbo's previous knowledge of this area quite pleasing.

https://www.reddit.com/r/StableDiffusion/comments/1s0carp/a_painter_with_50_years_of_figurative_work_just/

http://huggingface.co/datasets/Hafftka/michael-hafftka-catalog-raisonne