My ML Dev env

I use

Usually Apt and Nix.

- Pretty standard. Very productive setup!

 

Anaconda / conda > pip

  • conda has a package for Jupyter Lab

conda install conda-forge::jupyterlab

 

https://jupyter.org/

PyCharm + GitHub Copilot

  • PyCharm has good debugging options. You can use the JetBrains toolbox.

 

Neovide (on Linux Mint via Nix)

  • Doesn’t need all that configuration. Well suited for opening large files and jumping to a specific line number (:1234)

marius@mleng:~$ which nix /nix/var/nix/profiles/default/bin/nix marius@mleng:~$ nix-shell -p neovide ...

Linux Mint is rather conservative, which is good. As stable as Debian. Nix builds separate environments for modern packages like Neovide etc.