Does anyone know how to write a prompt for more than one person in a scene? I can’t get it to work.

I have been creating starter images using blender and a ‘make human’ plugin, and feeding them to img2img, but the first thing it does is to delete one of the people.

Maybe I need a different model or something.

Please advise, preferably with specifics.

    • chellebelle@lemmynsfw.com
      link
      fedilink
      English
      arrow-up
      5
      ·
      1 year ago

      Rendering at larger sizes can also make it more likely SD will generate multiple subjects. You can use the Regional Prompter or Latent Couple extensions for stable-diffusion-webui to control the subjects separately.

      You can also generate subjects separately using txt2img, paste them into a single scene, then do a pass with img2img to make the image look seamless. That is about the hardest approach, though. ControlNet with OpenPose and region prompter are my go-to solution.

      • Solated@lemmynsfw.com
        link
        fedilink
        English
        arrow-up
        2
        ·
        1 year ago

        Yeah I absolutely prefer the OpenPose approach because #1 you can guarantee that you will have (at least) the number of subjects that you want, it’s fairly straightforward, and I like having control over the pose.