Applies a painted realistic style to an image, similar to the classic Vampirella or Red Sonja comic books. Recommend starting prompt with "Digital illustration" and weighting the Lora between 0.4 and 0.6.
Have had very good results using the Photon model, RPG, Deliberate and Ares Mix. Anything that produces a well-formed human body should be usable, either realistic or not.
For V2, I more than doubled the number of training images, including additional artist styles from classic 1970's and 80's comics. More men included this time as well, so manly heroes and villains should render better.
Description
More than doubled the number of training images, including additional artist styles from classic 1970's and 80's comics. More men included this time as well, so manly heroes and villains should render better.
Continue to recommend weighting between .44 and .65. I prefer using realistic checkpoints such as Photon, but the Lora will work on illustration ones as well. Haven't tested much with anime checkpoints.