Installation¶
Glass Box UMAP is available on the Python Package Index (PyPI).
PyTorch Dependency
Glass Box UMAP requires PyTorch.
Existing Install: We respect and utilize any compatible version of PyTorch already found in your environment.
New Install: If PyTorch is missing, a compatible version will be automatically installed.
With uv¶
uv pip install glass-box-umap --torch-backend=auto
(If PyTorch is not found, installs version optimized for your hardware)
uv pip install glass-box-umap --torch-backend=auto
(If PyTorch is not found, installs version optimized for your hardware)
uv pip install glass-box-umap --torch-backend=auto
(If PyTorch is not found, installs version optimized for your hardware)
With pip¶
pip install glass-box-umap
(If PyTorch is not found, installs CUDA-enabled version)
pip install glass-box-umap
(If PyTorch is not found, installs CPU/MPS compatible version)
pip install glass-box-umap
(If PyTorch is not found, installs CPU-only version)
Warning
Using an NVIDIA GPU on Windows?
To use a CUDA-enabled version of PyTorch, install separately before installing Glass Box UMAP:
# Example for CUDA 12.4
pip install torch --index-url https://download.pytorch.org/whl/cu124
# Then install glass-box-umap
pip install glass-box-umap
From source¶
To install Glass Box UMAP from source, clone the repository and install it with uv:
git clone https://github.com/Arcadia-Science/glass-box-umap.git
cd glass-box-umap
uv sync
For developers/contributors, add the flags --extra plotting --group dev --group docs and see Contributing.