`syft.workers.tfe`¶

To be extended in the near future.

Module Contents¶

class syft.workers.tfe.TFEWorker(host=None, auto_managed=True)¶

start(self, player_name, cluster)¶: Start the worker as a player in the given cluster. Depending on whether the worker was constructed with a host or not this may launch a subprocess running a TensorFlow server.

stop(self)¶: Stop the worker. This will shutdown any TensorFlow server launched in start().

connect_to_model(self, input_shape, output_shape, cluster, sess=None)¶

Connect to a TF Encrypted model being served by the given cluster.

This must be done before querying the model.

query_model(self, data)¶

Encrypt data and sent it as input to the model being served.

This will block until a result is ready, and requires that a connection to the model has already been established via connect_to_model().

query_model_async(self, data)¶

Asynchronous version of query_model that will not block until a result is ready. Call query_model_join to retrive result.

This requires that a connection to the model has already been established via connect_to_model().

query_model_join(self)¶: Retrives the result from calling query_model_async, blocking until ready.

class syft.workers.tfe.TFECluster(*workers)¶

A TFECluster represents a group of TFEWorkers that are aware about each other and collectively perform an encrypted computation.