huggingface/pytorch_block_sparse

does this package allow multi-gpu training and distributed training?

DanqingZ opened this issue · 1 comments

If so, could you provide an example? Thanks.

Hi, I tried it with pytorch lightning using DDP and it works as expected