Encrypt both model weights and data for inference

[SanaAwan5]

Is there an equivalent command for encrypting the trained model weights similar to the encrypt_server_data=yes command for encrypting data before running inference?

I tried using the command line flag NGRAPH_ENCRYPT_MODEL =1 but it does not work for the latest version of HE_SEAL.

Thank you!

[fboemer]

We used to support encrypting model weights, but no longer do (https://github.com/NervanaSystems/he-transformer/commit/862e300f21d7b08b25ce5e878c1d18e6bb5c3414).