This Lora doesn't really do what I had wanted it to do, but the style is not unpleasant. Trained on the images of french comic artist Lewis Trondheim. The style sadly doesn't resemble his works when used at a strength of 1. You get closer to something resembling his character designs at 1.5, but then you get all kinds of visual errors especially in the faces and the background.
At a value between 1 and 1.2 you will get a comic book like style with clean black ink outlines, bright colors and simple shading.
Edit: Not sure what happened, but the generator suddenly does produce artwork very similar to Trondheim's style after a day has passed. I'll have to test more to be sure.
Description
FAQ
Comments (1)
You say it doesn't look how you wanted. How many images did you use? Handtagged or just auto-label?
I note a lot of your tags in the sample images ("very awa" for example) don't actually do anything, cleaning them up may help your image gens.
















