# tensorflow/gnn

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

Repository: https://github.com/tensorflow/gnn
Canonical: https://ross.abutalabs.com/products/gnn
Language: Python
License: Apache-2.0
License Family: permissive
Topics: tensorflow, gnn, deep-learning, machine-learning
Last push: 2026-07-31T09:30:11+00:00

## Health v2 (maintenance only)
Score: 66/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 95, release rhythm 8, longevity 100
- inputs: {"age_days": 1849, "days_push": 33, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1541, forks 203 (observed 2026-08-28T04:05:00.657892+00:00)

## What it is
TensorFlow GNN is a Python library for building Graph Neural Networks on TensorFlow, including a GraphTensor type for heterogeneous graphs, graph sampling tools, and ready-to-use GNN models. It is an open-source port of a Google-internal library used for scalable graph learning.

## Use cases
- train graph neural networks on heterogeneous graphs
- classify molecular graphs with GNNs
- sample subgraphs from a large graph database for training
- run node classification on citation networks like OGBN-MAG
- predict shortest paths with graph networks
- build custom GNN models with Keras layers

## When to choose
- you want to build GNNs within the TensorFlow/Keras 2 ecosystem
- you need to model heterogeneous graphs with multiple node and edge types
- you need scalable graph sampling from large graph databases
- you want ready-made GNN architectures plus a training runner API

## When to avoid
- your project uses Keras 3 or multi-backend Keras
- you prefer PyTorch-based graph learning (consider PyG or DGL instead)
- your data is not naturally graph-structured
- you need a framework-agnostic GNN solution

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, etl
- domain: machine-learning, deep-learning, data-science
- platform: python
- tags: graph-neural-networks, tensorflow, keras, heterogeneous-graphs, graph-sampling, gpu

## Member repositories
- tensorflow/gnn (main) score 66

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:00.657892+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 2026-08-30T04:30:48.950619+00:00, confidence not recorded.
  - readme: https://github.com/tensorflow/gnn (fetched 2026-08-28T04:05:00.657892+00:00, sha c878a724cb4e)
- Data as of 2026-08-30T08:39:29.467469+00:00.