pytorch

1a051046 - [PyTorch] Avoid extra Tensor refcounting in _cat_out_cpu (#49364)

[PyTorch] Avoid extra Tensor refcounting in _cat_out_cpu (#49364) Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/49364 We had a local `Tensor` when we only needed a `const Tensor&`. ghstack-source-id: 118624595 Test Plan: Internal benchmark. Reviewed By: hlu1 Differential Revision: D25544731 fbshipit-source-id: 7b9656d0371ab65a6313cb0ad4aa1df707884c1c