PyEnv

pyenv is a tool to manage python versions. It has a plugin called pyenv-virtualenv which manages virtual environments.

The documentation on the GitHub pages is really nice and succinct:

• pyenv Github
• pyenv-virtualenv Github

Shell Setup

I didn't want to use ~/.pyenv for my installation since it would make it messy but VSCode is annoying and didn't want to let me select the directory. So, I'm just going to move it to its default location so that I don't have to fight with VSCode.

# Pyenv integration
export PYENV_ROOT="$HOME/.pyenv"
[[ -d $PYENV_ROOT/bin ]] && export PATH="$PYENV_ROOT/bin:$PATH"
eval "$(pyenv init - zsh)"
eval "$(pyenv virtualenv-init -)"

I also ran pyenv global system so that it uses my global install of python when running globally.

Jupyter Environment Setup

I'm just going to show you commands for my machine-learning environment:

pyenv install 3.12  # because torch-vision doesn't support 3.13 yet.

# create an environment `machine-learning` with python version 3.12
pyenv virtualenv 3.12 machine-learning

# if ever in this directory or sub-directories, use the machine-learning env
# this is done by setting a file .python-version
pyenv local machine-learning

# verify
pyenv which python
pyenv which pip
python --version
pip --version

Now that the environment is setup, I had a requirements.txt from my previous machine learning environment which I just then ran pip -r requirements.txt

Now we want to installed ipykernel so that we can use the jupyter environment.

pip install ipykernel

Now we can use jupyter environment in VSCode. You can read more details about this on the VSCode Website.