How does cuGraph support GNNs?


Thanks for hosting this - sorry if this is a dumb question but I am interested to know how exactly does cuGraph support GNN’s

Thanks again for your attention.

cuGraph offers scalable performance on various graph algorithms up to trillions of edges. Through the cuGraph-DGL and cuGraph-PyG packages, cuGraph can serve as a backend to DGL or PyG with minimal code change. In addition, cuGraph-ops models (i.e. CuGraphSAGEConv, CuGraphGATConv) have been integrated directly into DGL and PyG and offer faster performance than the base DGL and PyG models. cuGraph can also accelerate graph sampling, allowing training of massive-scale graphs across multiple nodes and multiple GPUs.

Thanks !

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.