• LemmysMum@lemmy.worldOP
      link
      fedilink
      English
      arrow-up
      3
      arrow-down
      1
      ·
      8 months ago

      Appreciated. Trying to work that natural beauty angle. Glory in all it’s imperfections.

  • pron42@lemmynsfw.com
    link
    fedilink
    English
    arrow-up
    6
    ·
    8 months ago

    Can I ask what you’re using to get the realistic genitalia and consistent pose? Is this from a Lora?

    • LemmysMum@lemmy.worldOP
      link
      fedilink
      English
      arrow-up
      2
      arrow-down
      1
      ·
      edit-2
      8 months ago

      The pose is mostly just prompting, but in combination with a hairy pussy lora and number of adjustment lora’s (age, breast size, breast shape).

      For pose consistency you have to keep in mind what is and isn’t going to be visible. For example, if you use a lora that enhances breasts or is trained for vagina’s then those things will always be visible, so then you can use those to limit what poses you’ll get. If you prompt for breasts you’ll be less likely to get people facing away from the camera, you won’t get back shots, use vagina lora’s and that’s likely to be the focus without extra prompting.

      If you’re using an offline version of Stable Diffusion you can also use control net extensions to allow the use of depth maps and wire framing to set poses in pre-processing.

  • Uriel238@lemmynsfw.com
    link
    fedilink
    English
    arrow-up
    5
    ·
    8 months ago

    Bikini trims rather than full Brazilian. I assume this was what you were going for, rather than full bush.

    Though it would be interesting if the software concluded pubes came with natural shave lines.

    Still, well done!

    • LemmysMum@lemmy.worldOP
      link
      fedilink
      English
      arrow-up
      4
      arrow-down
      1
      ·
      edit-2
      8 months ago

      It’s variable. I was just throwing a few variations in but with more specific prompting you can probably get more specific with coverage, density, etc.

      Also it depends on the training data, this model seems to lack curly pubes with most of the model data obviously coming from Asian women.