no code implementations • 30 Oct 2021 • Shenghao Qiu, You Liang, Zheng Wang
Our model is first trained offline using training matrix samples, and the trained model can be applied to any input matrix and GNN kernels with SpMM computation.