• 0 Posts
  • 25 Comments
Joined 2 years ago
cake
Cake day: July 4th, 2023

help-circle

  • I know you said consumer GPU, but I run a used Tesla P40. It has 24 GB of vram. The price has gone up since I got it a couple years ago, there might be better options in the same price category. Still, it’s going to be cheaper than a modern full fat consumer gpu, with a reasonable performance hit.

    My use case is text generation, chat kind of things. In most cases, the inference is more than fast enough, but it can get slow when swapping out large context lengths.

    Mostly I run quantized 8-20B models with the sweet spot being around 12. For specialized use cases outside of general language, you can run more compact models. The general output is quite good, and I would have never had thought it was possible 10 years ago.

    ETA: I paid about $200 USD for the P40 a couple years ago, plus the price for a fan and 3d printed shroud.


  • I would do FDE yeah. My current laptop setup is with systemd-boot and a special initramfs that allows me to unlock it with a yubikey, with fallback to password. Fair warning, this exact configuration is not particularly easy to setup.

    There are also modules which enable early network connectivity along with a SSH server, meaning you login and unlock it remotely. I have not tried this.

    Debian does not frequently require rebooting under normal circumstances. Kernel updates are not that frequent, and you can usually put it off for a bit if you don’t want to deal with it.





  • Thought I’d throw in some things I’ve been using.

    • yetCalc - it’s a calculator
    • Vector Camera - novelty camera app which can make it look like you’re in the matrix
    • Leon - The URL Cleaner
    • Libera Reader - document viewer
    • K-9 Mail - it’s thunderbird mobile now
    • Imagepipe - image editor / exif cleaner
    • Gauguin - sudoku like puzzle game, configurable difficulty
    • Feeder - rss reader
    • AntennaPod - podcast manager / player
    • Element - flagship matrix client
    • AdAway - host based ad blocking (requires root)
    • Ente Auth - MFA




  • It sounds like you are having trouble with tonemapping HDR to SDR on the fly. This is a non-trivial task, but not impossible. Both mpv and ffmpeg (which plex and jellyfin use) are capable of this. If you install mpv, it will by default do the tonemapping, you can enable/disable this or force use of a particular algorithm if you like.

    To answer your question: Plex has been pretty shitty for years now, and it’s only getting worse. They just don’t care for their user base.

    ETA: Jellyfin also already does what you want, I think?



  • Gitlab uses a ton of resources and is a pain to setup. Once you get it going, it’s fine.

    Going to echo what others have said: Use Gitea or Forgejo instead if you can. Both have runners you can setup like gitlab, but they instead mimic github actions instead of gitlab ci/cd.

    I run a semi-private gitea instance, and have not had any problems past the initial setup in 2+ years.