[xnnpack][lite-int] preprocess (#86980)
Split up original preprocess diff:
This diff introduces the skeleton structure of the delegate APIs. first introducing the method compile spec error handling. For now it just outputs an empty tensor object upon execute. But just proves that delegate apis is working and a new xnnpack delegate backend has been added.
Differential Revision: [D38562918](https://our.internmc.facebook.com/intern/diff/D38562918/)
**NOTE FOR REVIEWERS**: This PR has internal Meta-specific changes or comments, please review them on [Phabricator](https://our.internmc.facebook.com/intern/diff/D38562918/)!
Pull Request resolved: https://github.com/pytorch/pytorch/pull/86980
Approved by: https://github.com/salilsdesai, https://github.com/cccclai
