pyg-lib

Installation

We provide pre-built Python wheels for all major OS/PyTorch/CUDA combinations, see here. Each wheel supports CPython 3.10 through 3.14 via CPython's stable ABI. CI checks every repaired wheel with abi3audit --strict and tests it on each supported CPython version.

To install the wheels, simply run

pip install pyg-lib -f https://data.pyg.org/whl/torch-${TORCH}+${CUDA}.html

where

• ${TORCH} should be replaced by either 2.12.0, 2.13.0, or 2.14.0
• ${CUDA} should be replaced by either cpu, cu130, or cu132

The following combinations are supported:

PyTorch 2.14	cpu	cu130	cu132
Linux	✅	✅	✅
Windows	✅	✅	✅
macOS	✅

PyTorch 2.13	cpu	cu130	cu132
Linux	✅	✅	✅
Windows	✅	✅	✅
macOS	✅

PyTorch 2.12	cpu	cu130	cu132
Linux	✅	✅	✅
Windows	✅	✅	✅
macOS	✅

From nightly

Nightly wheels are provided for Linux and support CPython 3.10 through 3.14:

pip install pyg-lib -f https://data.pyg.org/whl/nightly/torch-${TORCH}+${CUDA}.html

From source

pip install ninja wheel
pip install --no-build-isolation git+https://github.com/pyg-team/pyg-lib.git