Neural Networks

Linear(input_dims, output_dims[, bias])	Linear layer with Xavier Glorot weight inititalization.
MessagePassing([aggr])	Base class for creating Message Passing Neural Networks (MPNNs) [1].
GraphNetworkBlock([node_model, edge_model, ...])	Implements a generic Graph Network block as defined in [1].
SimpleConv([aggr, combine_root_func])	A simple non-trainable message passing layer.
GATConv(node_features_dim, out_features_dim)	Graph Attention Network convolution layer.
GCNConv(node_features_dim, out_features_dim)	Applies a GCN convolution over input node features.
GINConv(mlp[, eps, learn_eps, ...])	Graph Isomorphism Network convolution layer from "How Powerful are Graph Neural Networks?" paper.
SAGEConv(node_features_dim, out_features_dim)	GraphSAGE convolution layer from "Inductive Representation Learning on Large Graphs" paper.
GeneralizedRelationalConv(in_features_dim, ...)	Generalized relational convolution layer from Neural Bellman-Ford Networks: A General Graph Neural Network Framework for Link Prediction paper.
BatchNormalization(num_features[, eps, ...])	Applies batch normalization over a batch of features as described in the Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift" paper.
InstanceNormalization(num_features[, eps, ...])	Instance normalization over each individual example in batch of node features as described in the paper Instance Normalization: The Missing Ingredient for Fast Stylization paper.
LayerNormalization(num_features[, eps, ...])	Applies layer normalization over each individual example in a batch of features as described in the Layer Normalization paper.
global_add_pool(node_features[, batch_indices])	Sums all node features to obtain a global graph-level representation.
global_max_pool(node_features[, batch_indices])	Takes the feature-wise maximum value along all node features to obtain a global graph-level representation.
global_mean_pool(node_features[, batch_indices])	Takes the feature-wise mean value along all node features to obtain a global graph-level representation.