Download ML thing.
make new venv.
pip install -r requirements.txt.
pip can’t find the right versions.
pip install --update pip.
pip still can’t find the right versions.
install conda.
conda breaks for some reason.
fix conda.
install with conda.
pytorch won’t compile with CUDA support.
install 2,000,000GB of nvidia crap from conda.
pytorch still won’t compile.
install older version of gcc with conda.
pytorch still won’t compile.
reinstall the entire operating system with debian 11.
apt can’t find shitlib-1.
install shitlib-2.
it’s not compatible with shitlib-1.
compile it from source.
automake breaks.
install debian 10.
It actually works.
“Join our discord to get the model”.
give up.

I’ve recently discovered pipenv, and it has been a massive QoL improvement. No need to figure out bazillion of commands just to create or start an environment, or deal with what params should you use for it like you do with venv. You just pipenv install -r requirements.txt, and everything is handled for you. And when you need to run it, just pipenv run python script.py and you are good to go.

The best thing however are the .pipfiles, that can be distributed instead of requirements.txt, and I don’t get why it’s not more common. It’s basically requirements, but directly for pipenv, so you don’t need to install anything and just pipenv run from the same folder.

As a mac user I feel this

Poetry is nice for this but honestly, Rust’s cargo and the JS npm/yarn have spoilt me.

Hopefully Mojo will sort it all out. Maybe even inspiring a new, positive streak of xkcd strips in the future?

[email protected]