pytorch

201c36d8 - Hack get_nbytes() to return 0 for sparse tensors as workaround for functionalization (#90702)

Hack get_nbytes() to return 0 for sparse tensors as workaround for functionalization (#90702) Pull Request resolved: https://github.com/pytorch/pytorch/pull/90702 Approved by: https://github.com/ezyang, https://github.com/bdhirsh