A LoRA trained on images of hand drawn images. The only needed trigger is <code>traditional media</code>, yet I do recommend using <code>sketch</code> if you want to guarantee the results.

Traditional Media

Newer training method, more samples, should behave way better now. Trained on WAI-NSFW-illustrious-SDXL, so it will work better there, but should be usable with all Illustrious models.

Files with SHA256: db046291f91b0cf97e72da64b07506e666d8d8c6635fdf5f927225ee4908d3ec