GNN Cheatsheet

• SparseTensor: If checked (✓), supports message passing based on torch_sparse.SparseTensor, e.g., GCNConv(...).forward(x, adj_t). See here for the accompanying tutorial.

• edge_weight: If checked (✓), supports message passing with one-dimensional edge weight information, e.g., GraphConv(...).forward(x, edge_index, edge_weight).

• edge_attr: If checked (✓), supports message passing with multi-dimensional edge feature information, e.g., GINEConv(...).forward(x, edge_index, edge_attr).

• bipartite: If checked (✓), supports message passing in bipartite graphs with potentially different feature dimensionalities for source and destination nodes, e.g., SAGEConv(in_channels=(16, 32), out_channels=64).

• static: If checked (✓), supports message passing in static graphs, e.g., GCNConv(...).forward(x, edge_index) with x having shape [batch_size, num_nodes, in_channels].

• lazy: If checked (✓), supports lazy initialization of message passing layers, e.g., SAGEConv(in_channels=-1, out_channels=64).

Graph Neural Network Operators

Name	SparseTensor	edge_weight	edge_attr	bipartite	static	lazy
SimpleConv	✓	✓		✓	✓
GCNConv (Paper)	✓	✓			✓	✓
ChebConv (Paper)		✓			✓	✓
SAGEConv (Paper)	✓			✓	✓	✓
CuGraphSAGEConv (Paper)					✓
GraphConv (Paper)	✓	✓		✓	✓	✓
GatedGraphConv (Paper)	✓	✓			✓
ResGatedGraphConv (Paper)	✓		✓	✓	✓	✓
GATConv (Paper)	✓		✓	✓		✓
CuGraphGATConv (Paper)					✓
FusedGATConv (Paper)					✓
GATv2Conv (Paper)	✓		✓	✓		✓
TransformerConv (Paper)	✓		✓	✓		✓
AGNNConv (Paper)	✓				✓
TAGConv (Paper)	✓	✓			✓	✓
GINConv (Paper)	✓			✓	✓
GINEConv (Paper)	✓		✓	✓	✓
ARMAConv (Paper)	✓	✓			✓	✓
SGConv (Paper)	✓	✓			✓	✓
SSGConv (Paper)	✓	✓			✓	✓
APPNP (Paper)	✓	✓			✓
MFConv (Paper)	✓			✓	✓	✓
DNAConv (Paper)	✓	✓
GMMConv (Paper)	✓		✓	✓	✓	✓
SplineConv (Paper)	✓		✓	✓	✓	✓
NNConv (Paper)	✓		✓	✓	✓	✓
CGConv (Paper)	✓		✓	✓	✓
EdgeConv (Paper)	✓			✓	✓
FeaStConv (Paper)	✓			✓	✓	✓
LEConv (Paper)	✓	✓		✓	✓	✓
PNAConv (Paper)	✓		✓			✓
ClusterGCNConv (Paper)	✓				✓	✓
GENConv (Paper)	✓		✓	✓	✓	✓
GCN2Conv (Paper)	✓	✓			✓
PANConv (Paper)	✓				✓	✓
WLConv (Paper)	✓				✓
WLConvContinuous (Paper)	✓	✓		✓	✓
SuperGATConv (Paper)	✓					✓
FAConv (Paper)	✓	✓		✓	✓	✓
EGConv (Paper)	✓					✓
PDNConv (Paper)	✓		✓		✓
GeneralConv (Paper)	✓		✓	✓		✓
LGConv (Paper)	✓	✓			✓
GPSConv (Paper)	✓				✓
AntiSymmetricConv (Paper)	✓				✓
DirGNNConv (Paper)					✓
MixHopConv (Paper)	✓	✓			✓	✓

Heterogeneous Graph Neural Network Operators

Name	SparseTensor	edge_weight	edge_attr	bipartite	static	lazy
RGCNConv (Paper)	✓
FastRGCNConv	✓
CuGraphRGCNConv (Paper)					✓
RGATConv (Paper)	✓		✓
FiLMConv (Paper)	✓			✓	✓	✓
HGTConv (Paper)	✓					✓
HEATConv (Paper)	✓		✓			✓
HeteroConv					✓
HANConv (Paper)	✓					✓

Hypergraph Neural Network Operators

Name	SparseTensor	edge_weight	edge_attr	bipartite	static	lazy
HypergraphConv (Paper)		✓	✓			✓

Point Cloud Neural Network Operators

Name	bipartite	lazy
GravNetConv (Paper)	✓	✓
SignedConv (Paper)	✓	✓
PointNetConv (Paper)	✓
DynamicEdgeConv (Paper)	✓
XConv (Paper)
PPFConv (Paper)	✓
PointTransformerConv (Paper)	✓	✓
HeteroConv
PointGNNConv (Paper)