Initially created a dataset for a Flux model - tried training for SDXL and got nearly the same results! Stronger anatomy and prompt coherence with Flux, but still quite good here.The original <a target="_blank" rel="ugc" href="https://civarchive.com/models/262880/eldritch-comics-comic-book-style-illustration">Eldritch Comics</a> relies so heavily on its colorist. But for Eldritch Dark Comics, our virtual artist wants his black pen to take prominence.This initial version definitely has its flaws, particularly with hands. But I'm having fun exploring it right now. Let me know what other deficiencies you identify and I'll see if they can be improved in another iteration.

Eldritch Dark Comics | for SDXL

initial version - same training dataset as for the first-trained Flux.1 version

Files with SHA256: 01ed539287d73db86aaa62ed45f445e45f0ef337cd9b5264355dab01a73bbede