Poor nVidia Jetson, you did great for the last 5 years.

Managed to fry the eMMC by shorting pins, it looks like.

Note for future self: fully enclose boards with tight spaces.

  • recklessengagement@lemmy.world
    link
    fedilink
    arrow-up
    0
    ·
    6 months ago

    Ahh gotcha, similar to Google’s Coral? Neat.

    I’ve recently been looking into locally hosting some LLMs for various purposes, I haven’t specced out hardware yet. Any good resources you can recommend?

    • FooBarrington@lemmy.world
      link
      fedilink
      arrow-up
      1
      ·
      6 months ago

      Ahh gotcha, similar to Google’s Coral?

      Kind of, it’s a standalone system with the hardware integrated - kinda like Google Coral with a Raspberry Pi.

      I’ve recently been looking into locally hosting some LLMs for various purposes, I haven’t specced out hardware yet. Any good resources you can recommend?

      Not really, sorry - I haven’t gone too deep into LLMs beyond simple use cases. I’ve only really used llama.cpp myself.

    • Linkerbaan@lemmy.world
      link
      fedilink
      arrow-up
      0
      arrow-down
      1
      ·
      edit-2
      6 months ago

      A dedicated NVIDIA GPU in a random x86 pc is a lot faster and more price efficient than a Jetson.

      If it isn’t about the form factor the Jetson is not a great contender.