Efficient Message Passing Architecture for GCN Training on HBM-based FPGAs with Orthogonal Topology On-Chip Networks

Qizhe Wu, Letian Zhao, Yuchen Gui +1

Graph Convolutional Networks (GCNs) are state-of-the-art deep learning models for representation learning on graphs. However, the efficient training of GCNs is hampered by constrai…