Installation

Create and activate an environment, then install the package in editable mode.

python3 -m venv .venv
source .venv/bin/activate
python -m pip install -U pip
python -m pip install -e ".[data,notebooks,docs,dev]"

Install framework extras as needed:

python -m pip install -e ".[tensorflow]"
python -m pip install -e ".[pytorch]"
python -m pip install -e ".[jax]"

For NVIDIA GPU runs on Linux:

python -m pip install -e ".[tensorflow,pytorch,jax-cuda13]"

The TensorFlow CLI automatically restarts once with the pip-installed NVIDIA libraries on LD_LIBRARY_PATH and XLA_FLAGS pointing at CUDA libdevice. For notebooks, start Jupyter from a shell where those variables are already available, or run the CLI for full GPU reproductions.

For everything in one environment:

python -m pip install -e ".[all]"

For package build checks without uploading to PyPI:

python -m pip install -e ".[release]"
python -m build
python -m twine check dist/*

The processed .npy files are kept in Git LFS for repository reproducibility, but they are excluded from the PyPI source distribution and wheel.