The aim of this keras extension is to provide Sequential and Functional API for performing deep learning tasks on graphs. Specifically, Keras-DGL provides implementation for these particular type of layers,
- Graph Convolutional Neural Networks (GraphCNN).
- Graph Attention Convolutional Neural Networks (GraphAttentionCNN).
- Graph Convolutional Recurrent Neural Networks (GraphConvLSTM).
- Graph Capsule Convolutional Recurrent Neural Networks (GraphCapsuleCNN) TBD.
- Graph Message Passing Neural Networks (GraphNeuralNetworks) TBD.
- Keras-DGL also contains implementation of various graph convolutional filters TBD.
Read the documentation: http://vermaMachineLearning.github.io/keras-deep-graph-learning