CogDL: A Comprehensive Library for Graph Deep Learning

Published in ACM Web Conference (WWW), 2023

CogDL provides a unified design for training and evaluation of GNN models across graph tasks, with efficient sparse operators and support for mixed precision training.

Recommended citation: Cen, Y., Hou, Z., Wang, Y., Chen, Q., Luo, Y., Yu, Z., et al. (2023). "CogDL: A Comprehensive Library for Graph Deep Learning." WWW.
Download Paper

GitHub