Seems like a really cool project. Lowering the barrier to entry of locally run models. As llamacpp supports a ton of models, I imagine it be easy to adapt this for other models other than the prebuilt ones.

  • kakes@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    14
    arrow-down
    1
    ·
    edit-2
    1 year ago

    My first thought is that this makes it a lot easier for a game dev to create a game that uses a local LLM. I’ll definitely be keeping an eye on this for my own projects.

    Edit: Trying it out now, color me impressed! I opened the file, and it worked instantly without a single finick to be found. Super excited for this!

    • kakes@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      1
      ·
      1 year ago

      When I upload an image, it seems to be getting re-encoded every time I send a message, which takes a while. I’ve never really messed around with images in Llama - is this a normal/necessary thing?