This is a LoRA (XL) for creating VEDALKEN.

Keep the model contribution relatively high (like 0.70 or greater). The bias that the checkpoints have for just plopping ears on the model is pretty strong even though all of this model's training data does not have ears.

Anyhow, the RealCartoonXL checkpoint really showed some great results, but several others were fine, too.

I do need to do another version that really focuses on no ears, but I strongly recommend using (((ears))) in the negative prompt.

The trigger words for the model are as follows:

1boy, 1girl, abs, armlet, armor, bald, belt, blue eyes, blue skin, blurry, blurry background, bodysuit, book, breasts, brown eyes, cape, cleavage, closed eyes, closed mouth, colored skin, earrings, facial hair, facial mark, facing viewer, gloves, glowing, glowing eyes, gorget, helmet, hood, jacket, jewelry, lips, looking at viewer, male focus, mask, medium breasts, midriff, muscular, navel, necklace, nipples, nose, portrait, purple eyes, realistic, red eyes, ring, robe, shoulder armor, singl