awslabs / dgl-ke

High performance, easy-to-use, and scalable package for learning large-scale knowledge graph embeddings.
https://dglke.dgl.ai/doc/
Apache License 2.0
1.25k stars 194 forks source link

RuntimeError: Cannot re-initialize CUDA in forked subprocess #271

Open avi-pomicell opened 1 year ago

avi-pomicell commented 1 year ago

Hi

I'm getting this error when trying to use dgl-ke with GPU: RuntimeError: Cannot re-initialize CUDA in forked subprocess. To use CUDA with multiprocessing, you must use the 'spawn' start method

I am able to run the code on CPU only. Any ideas?

More info: I set up the environment using these insturctions: https://github.com/awslabs/dgl-ke/issues/267#issuecomment-1347933918

XinliYu commented 11 months ago

Same issue here.