Moss can natively generate comics and image sequences of the same character in any style
retained text
Moss SFT int8 that I merged with the stripe fix lora + Qwen 9b int8 So appearently it can create story sequences or comics within the same image. This model is insane. It follows prompts extremely well down to the smallest details. Manual simple sentence prompting. I used Lenovo lora for the photo style images. Generated with my experimental turbo method with the dmd2 turbo lora but high cfg to fix blur and bokeh problems with negative prompts so it was still quite fast. All pics are native 1080p. This model has huge potential for image editing and editing loras.
document fetched 2026-10-24 10:34 · canonical https://reddit.com/r/StableDiffusion/comments/1xdnxkr/moss_can_natively_generate_comics_and_image