Label Propagation out of memory

I want to run Label propagation algorithms from dgl on GPU. My graph just have 1 node (customer), the statistic of the graph is more then 5 millions node and 9 millions edges. When I run graph for label propagation function, it raise an error of out of memory. Is there any way that I can fix this problem?
Thank yiu!

Hi @minhntn, DGL supports place the graph and features in the CPU memory while computing in the GPU. See uva in dgl dataloader.

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.