gnn-tf
A graph neural network framework on tensorflow.
Developed by
License
Apache License 2.0 (Apache-2.0)
Main Characteristic
Provides a framework for easy experimentation with Graph Neural Network (GNN) architectures by separating them from predictive components.
- GNNs
- Rapid prototyping
- Modular architectures
- Predictive tasks independent of architectures
- Common training and losses
- Use tensorflow code (e.g. activations, optimizers, keras layers)
- Easy layer injection/removal
Last updated
10.11.2022 - 09:47