Just training some loras from the data I used for SD1.5 training.

This
was supposed to be 'pencil art', but Anima prefers to color images it
outputs. Tag itself didn't take in any monochrome line art stylization.

What
it did take in, is overall muscular chad look dominating the dataset.
It's also seemingly a little better with various art medium textures.

If you want line art you will have to force it a little - 'sketch, line art, monochrome', maybe remove negative prompt.