How to use HGT to build models and classifiers to implement graph classification of heterogeneous graphs? thanks.
You may refer to this example: dgl/examples/pytorch/hgt at master · dmlc/dgl · GitHub to get the node embeddings for each node.
For graph classification, a typical approach will be pooling (e.g. taking an average) of all the node embeddings in a graph, then apply an MLP afterwards.
Thank you so much for your assistance.