[PT-D][Sharding] Clean up sharded tensor code by leverage handle_torch_function (#76824)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/76824
In https://github.com/pytorch/pytorch/pull/75374, we have added some ops for sharded tensor and there are some duplicate code which can be removed by `leveraging handle_torch_function`
ghstack-source-id: 155503715
Test Plan: CI
Reviewed By: wanchaol
Differential Revision: D36135157
fbshipit-source-id: 31e854a69c68a7e938ef9d6d5c71ff587af1d7aa
(cherry picked from commit af2bb172f66a0808d3932d60e404b989bb41d1bc)