Enable typechecks for torch.nn.quantized.modules.linear (#44154)
Summary:
Also import `Optional` directly from `typing` rather than from `_jit_internal`
Pull Request resolved: https://github.com/pytorch/pytorch/pull/44154
Reviewed By: seemethere
Differential Revision: D23511833
Pulled By: malfet
fbshipit-source-id: f78c5fd679c002b218e4d287a9e56fa198171981
