I am looking for Linux distro for my workstation. My daily job consist of different data science and machine learning tasks. Which Linux workstation distro is considered good for data science and machine learning? Is there any general consensus? What is your personal experience doing data science and machine learning on Linux workstation?

  • cyril1991@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Conda environments, containers and portability > in-depth Linux distro configuration. You really want to keep the distro configuration to a minimum, ultimately it is a distraction (“cattle not pets”).

  • BetterComment@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    I started with Ubuntu and saw no reason to use anything else, so I would just use 1 and stop wasting your time/energy thinking about it

  • Automatic-Cycle-9891@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Ubuntu is great. When I started looking into machine learning 4 years ago, it became real clear to stay away from windows. At the time, Ubuntu was a dominant distro so I took the plunge and never looked back.

    Today I had to boot up into Windows and was once again reminded of why Ubuntu rocks. Windows just gets in the way …

  • BetterComment@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    I started with Ubuntu and saw no reason to use anything else, so I would just use 1 and stop wasting your time/energy thinking about it

  • cyril1991@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Conda environments, containers and portability > in-depth Linux distro configuration. You really want to keep the distro configuration to a minimum, ultimately it is a distraction (“cattle not pets”).

  • brianplusplus@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Pop!_os, or arch, they are both equal for ML but i can sure tell you which one is easier to set up…

  • robml@alien.top
    cake
    B
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Honestly:

    • Mint if completely new
    • Endeavor if you have typed on the terminal a bit before
    • Debian or Arch for more comfortable users: the more preferring slow changes the latter preferring more, however both are STABLE so up to you and your use case.
  • brianplusplus@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Pop!_os, or arch, they are both equal for ML but i can sure tell you which one is easier to set up…

  • robml@alien.top
    cake
    B
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Honestly:

    • Mint if completely new
    • Endeavor if you have typed on the terminal a bit before
    • Debian or Arch for more comfortable users: the more preferring slow changes the latter preferring more, however both are STABLE so up to you and your use case.
  • Automatic-Cycle-9891@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Ubuntu is great. When I started looking into machine learning 4 years ago, it became real clear to stay away from windows. At the time, Ubuntu was a dominant distro so I took the plunge and never looked back.

    Today I had to boot up into Windows and was once again reminded of why Ubuntu rocks. Windows just gets in the way …

  • mfarahmand98@alien.topB
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Manjaro. The repository has an excellent GUI and it has everything you’ll ever need (AUR included).