C++

pyg-lib

Low-Level Graph Neural Network Operators for PyG

P

pyg-team

Dernière activité 28 sept. 2026
pyg-team/pyg-lib

239

étoiles

64

forks

27

issues ouvertes

deep-learninggeometric-deep-learninggraph-convolutional-networksgraph-neural-networkspytorch

Ce README est souvent en anglais.

pyg-lib

Docs Status Code Coverage Slack Contributing

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

Projets similaires

Graph Neural Network Library for PyTorch

Pythondeep-learninggeometric-deep-learninggraph-convolutional-networks
Ppyg-team
24,1 k étoiles4,1 k

Tensors and Dynamic neural networks in Python with strong GPU acceleration

Pythonautograddeep-learninggpu
Ppytorch
103,5 k étoiles30,9 k

TensorFlow GNN is a library to build Graph Neural Networks on the TensorFlow platform.

Pythondeep-learninggnnmachine-learning
Ttensorflow
1,5 k étoiles205