pytorch

6f0e5362 - Enforce that named_tensor_meta_ is non-null only if there is a non-wildcard name (#38725)

Enforce that named_tensor_meta_ is non-null only if there is a non-wildcard name (#38725) Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/38725 Today, there are two equivalent representations: named_tensor_meta_ is null, or named_tensor_meta_ is non-null but all of the dimension names are wildcard. Let's reduce the opportunity for behavior divergence by making the second representation illegal. This will make it easier for me to add a dispatch key for named tensor as I can rely on setters to always go through TensorImpl to maintain invariants on DispatchKey. Signed-off-by: Edward Z. Yang <ezyang@fb.com> Test Plan: Imported from OSS Differential Revision: D21662641 Pulled By: ezyang fbshipit-source-id: ccc6566d23ad2ba850f653364a86cc8db0428223