A graph neural network built with TensorFlow2.
Graph Neural Network (GNN)
Graph Neural Network library written in TensorFlow 2 and roughly based on the work of Battaglia et al..
Supported layer types:
- Dense layer with dropout
- Probabilistic layer (planned)
First, install the virtual python environment by running
bash install.sh. Before executing the bash scripts, make sure that these are executable using
chmod -x *.sh. You can enter the virtual environment using
Once, the basics are set-up, run
bazel test ... to make sure all is up and running.
In the example provided, the GNN learns the costs of nodes based on the sum of distances to three neighboring nodes.
To run this example use this command:
bazel run //examples:cost_based_on_dist.
Figure 1: Graph shows the real (c) and the predicted costs by the GNN (p).
This library is an open source project available on Github and I cordially invite everyone to contribute to this project.
Copyright (c) 2019 Patrick Hart
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size||File type||Python version||Upload date||Hashes|
|Filename, size tf_gnn-0.0.1-py3-none-any.whl (11.3 kB)||File type Wheel||Python version py3||Upload date||Hashes View hashes|
|Filename, size tf-gnn-0.0.1.tar.gz (6.6 kB)||File type Source||Python version None||Upload date||Hashes View hashes|