Model decryption prior to loading

[vadimkantorov]

I found that support for a custom user hook for model decryption was added to TRT EP:

• [TRT EP] Fix logic to reach cache encryption code. #17111

Is similar functionality supported for other EPs (and primarily the default ORT cuda backend)?

Are there anywhere examples of using this model decryption feature?

Thanks!