Enable BF16 for indexing on CUDA (#48801)
Summary:
Fixes #{issue number}
Pull Request resolved: https://github.com/pytorch/pytorch/pull/48801
Reviewed By: glaringlee
Differential Revision: D25542914
Pulled By: ngimel
fbshipit-source-id: 4113eb2729d15b40a89268172cc37122b5213624