Open yuanqing-wang opened 4 years ago
we decided that for now we are going to pass only features in our message passing framework. this is safer but might add complexity in the future when we also wanted to have heterograph
looks like it's recommended by dgl to pass only tensors around and we haven't yet find a way to efficiently copy the graph.
dgl
note that calling g.local_var() doesn't copy the graph
g.local_var()
we decided that for now we are going to pass only features in our message passing framework. this is safer but might add complexity in the future when we also wanted to have heterograph