# pyg-team/pytorch_geometric

Graph Neural Network Library for PyTorch

Repository: https://github.com/pyg-team/pytorch_geometric
Canonical: https://ross.abutalabs.com/products/pytorch_geometric
Homepage: https://pyg.org
Language: Python
License: MIT
License Family: permissive
Topics: pytorch, geometric-deep-learning, graph-neural-networks, deep-learning, graph-convolutional-networks
Last push: 2026-08-24T20:53:29+00:00

## Health v2 (maintenance only)
Score: 84/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 55, longevity 100
- inputs: {"age_days": 3253, "days_push": 9, "days_rel": 89, "gap_med": 233, "n_releases_24m": 4}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 24032, forks 4039 (observed 2026-08-28T04:11:35.513111+00:00)

## What it is
PyG (PyTorch Geometric) is a Python library built on PyTorch for writing and training Graph Neural Networks on structured and irregular data. It provides GNN layers from published research, mini-batch loaders, benchmark datasets, and transforms for graphs, 3D meshes, and point clouds.

## Use cases
- train graph neural networks for node classification
- build GNN models for link prediction
- run deep learning on 3D meshes and point clouds
- load and mini-batch many small graphs or one giant graph
- experiment with graph convolutional network architectures
- benchmark GNN models on standard graph datasets

## When to choose
- you use PyTorch and need graph deep learning with a tensor-centric API
- you want many pre-implemented GNN models from research papers
- you need multi-GPU training and torch.compile support for GNNs
- you work with irregular structures like graphs, meshes, or point clouds

## When to avoid
- your project is built on TensorFlow or JAX rather than PyTorch
- you only need classical graph algorithms without deep learning
- you need a no-dependency lightweight graph library

## Facets
- artifact type: library
- maturity: stable
- function: deep-learning, machine-learning, data-science
- domain: machine-learning, deep-learning
- platform: python, cross-platform
- tags: graph-neural-networks, geometric-deep-learning, pytorch, gnn, graph-convolutional-networks, point-clouds, 3d-meshes, algorithms, gpu

## Member repositories
- pyg-team/pytorch_geometric (main) score 84

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:35.513111+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:56:17.972916+00:00, confidence not recorded.
  - readme: https://github.com/pyg-team/pytorch_geometric (fetched 2026-08-28T04:11:35.513111+00:00, sha f4c703fa4dfe)
- Data as of 2026-08-30T08:39:29.467469+00:00.