Remove output_spec from AOTInductorModelCache (#110462)
Summary:
No need to store output_spec as the returned exported.call_spec already contains that.
X-link: https://github.com/pytorch/pytorch/pull/110462
Approved by: https://github.com/angelayi
Reviewed By: PaliC
Differential Revision: D49891916
Pulled By: desertfire
fbshipit-source-id: 2501412a1b9b6c637ec24ce00d5fb673b569e778
