>>26854

ANIMA is capable of generating mild R-18G content using natural language prompts without the need for LoRAs, though the visual detail is somewhat lacking. My usual workflow, therefore, involves creating a rough sketch with ANIMA, upscaling and enhancing it, and then switching to an Illustrious-based model—such as WAI V17—for post-processing and detail refinement.

For example, this is an image generated directly by ANIMA; the composition matches my vision, but the details are poor.