Create helpers for implementing unary ops whose CUDA implementation is ATen. (#24879)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/24879
Pull Request resolved: https://github.com/pytorch/pytorch/pull/24879
Test Plan: Imported from OSS
Differential Revision: D17073557
Pulled By: VitalyFedyunin
fbshipit-source-id: 0d876627d500601ecd2a6aa6501e880842f2e98b