Streamlined interface for generating images with AI in Krita. Inpaint and outpaint with optional text prompt, no tweaking required. - GitHub - Acly/krita-ai-diffusion: Streamlined interface for gen...
  • Contend6248@feddit.de
    131·
    1 year ago

    Cool project!

    Would love AMD Linux support though, Stable Diffusion is fairly easy to install, but i just don’t like the Stable Diffusion WebUI

    A Flatpak would be golden

  • Kaldo@kbin.social
    5·
    1 year ago

    Am I understanding correctly and this is truly FOSS and fully offline, there’s no remote server or model we have to connect to? What was the model trained on? I’m really curious but I also don’t want to support proprietary unethical data sourcing.

    • Zerush@lemmy.mlOP
      2·
      1 year ago

      It create a local server, in the setup (~10GB disk), as say, selfhosted.

      • Otter@lemmy.caEnglish
        2·
        1 year ago

        What would be the minimum requirements outside of the disk space?

        I assume I’m hitting a limit, but I’m not sure what kind of upgrade I’d need to use it effectively

        • Zerush@lemmy.mlOP
          2·
          1 year ago

          I remember reading something on GitHub. I think it should work on a current PC/Laptop, the question is what resolution you can render and how long it will last.

  • Otter@lemmy.caEnglish
    5·
    1 year ago

    Giving it a try now! Haven’t tried Krita before either.

    Edit: the server seems to error out when I try to run it, testing with the CPU version instead. Just need to wait for it to finish

  • Helix 🧬@feddit.deEnglish
    4·
    1 year ago

    This is very cool, sadly AMD GPUs in Linux are not supported. I hope the server component can be run independently there at some point.

    • Wes_Dev@lemmy.ml
      2·
      1 year ago

      I’ve got an RX 580 8GB. What sucks is that it USED TO SUPPORT running AI and stuff. But AMD removed that support in new versions of the driver. Might have had a good reason, might not have. Still, sucks.