Config determines how a session should run in tf-encrypted.
There are two primary ways in which config can be used:
class tf_encrypted.config.LocalConfig


class tf_encrypted.config.RemoteConfig

As the name implies, LocalConfig is used to create a local session. This is useful for quick debugging and prototyping.

RemoteConfig is more robust and is used to specify how a graph will run in a production environment. What machines are on which host, etc.

See class definitions for usage examples and more.