Thanks for the great collection!!
We would like to bring to your attention our ICLR 2020 paper "GraphSAINT: Graph Sampling Based Inductive Learning Method". We propose a new minibatch training framework for general GNN models (e.g., GraphSAGE, GAT, JK-Net, MixHop, etc), which significantly improves the training efficiency and quality for large graphs and deep models.
Our code is also available at https://github.com/GraphSAINT/GraphSAINT
Thanks for your consideration.
Thanks for the great collection!!
We would like to bring to your attention our ICLR 2020 paper "GraphSAINT: Graph Sampling Based Inductive Learning Method". We propose a new minibatch training framework for general GNN models (e.g., GraphSAGE, GAT, JK-Net, MixHop, etc), which significantly improves the training efficiency and quality for large graphs and deep models.
Our code is also available at https://github.com/GraphSAINT/GraphSAINT
Thanks for your consideration.