OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Developer Tools & SDKs

Gnn

Gnn, also named TensorFlow GNN, is an open-source Python library for building graph neural networks on TensorFlow. It provides GraphTensor representations for graphs with multiple node and edge types, data-preparation tools including subgraph sampling, ready-made models and Keras layers, and a high-level API for training orchestration.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Gnn — OpenViST project profile in AI Developer Tools & SDKs.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use TensorFlow GNN when a TensorFlow graph-learning project needs graph representations, sampled subgraph preparation, Keras modeling components or training orchestration.

Who it is for

  • Machine-learning developers building graph neural networks with TensorFlow.
  • Teams preparing sampled subgraphs from large graph datasets for training or inference.
  • Researchers exploring graph tasks such as molecular classification or shortest-path prediction.

Documented uses

  • Represent heterogeneous graphs with multiple types of nodes and edges.
  • Prepare streams of smaller subgraphs from large graph data.
  • Build graph-learning models with supplied models and Keras layers.
  • Orchestrate graph neural network training.
Before you choose

Requirements & limits

The repository provides a pip installation command and browser-based Colab examples that require no local installation. Apache-2.0 licensed open-source code is available in the public repository.

  • TensorFlow 2.12 or higher and Keras v2 are required; TensorFlow 2.16+ also requires tf-keras and the TF_USE_LEGACY_KERAS=1 setting.
  • Apache Beam is required for distributed graph sampling.
  • The project is developed and tested on Linux; use on other TensorFlow-supported platforms is described as possible, not assured.
  • Colab examples cover molecular graph classification, heterogeneous OGBN-MAG training and shortest-path prediction.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Install TensorFlow GNN in a TensorFlow environment

Install the stable TensorFlow GNN package after preparing its documented TensorFlow and Keras requirements.

Before you start

  • Python package installation with pip.
  • TensorFlow 2.12 or higher and Keras v2.
  • For TensorFlow 2.16 or higher, install tf-keras and set TF_USE_LEGACY_KERAS=1.

Steps

  1. Prepare an environment with TensorFlow 2.12 or higher and Keras v2.

  2. Install the package with pip install tensorflow-gnn.

  3. If using TensorFlow 2.16 or higher, install tf-keras and set TF_USE_LEGACY_KERAS=1.

Expected outcome

The TensorFlow GNN package is installed in the prepared environment.

Guide limitations

  • TensorFlow GNN does not work with multi-backend Keras v3.
  • Apache Beam is needed for distributed graph sampling.
  • The project is developed and tested on Linux; other TensorFlow-supported platforms may work.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
tensorflow/gnn · Official homepage title · retained in its source language
How we review sourcesSuggest a correction