Used by the orchestrator to compute on the shares. Arguments: session (Session): the session secret (Optional[Union[torch.Tensor,

Class used to keep information about computation done in SMPC. Attributes: id (UID): The id to store the session tags (Optional[L

Falcon Protocol Implementation.

Session configuration. Attributes: encoder_base (int): Base for the encoder. encoder_precision (int): Precision for the encoder.

RSTensor is used when a party holds more than a single share,required by various protocols. Arguments: shares (Optional[List[Union[float,

Single Share representation. Arguments: data (Optional[Any]): the share a party holds session (Optional[Any]): the session from w

Encoding/decoding a tensor to/from a fixed precision representation. This class was inspired from the Facebook Research - CrypTen project At

Raised when there is no primitive in the store to do the computation.

ABY3 Protocol Implementation.

Manages items needed for MPC Computation. Attributes: store (Dict[Any, Any]): Object needed for MPC.

Function Secret Sharing.

Simple convolutional network.

A trusted third party should use this class to generate crypto primitives.

Class used to manage sessions. Attributes: uuid (Optional[UUID]): used to identify a session

Convolutional 2D.

The addition gradient function.

The multiplication gradient function.

The Division gradient function.

The Flatten gradient function.

Abstract class that should be implemented by the Gradient Functions.

The multiplication gradient function.

The MaxPool2D gradient function.

The multiplication gradient function.

The multiplication gradient function.

The sigmoid gradient function.

The Reshape gradient function.

The sigmoid gradient function.

The substraction gradient function.

The summation gradient function.

The tranpose gradient function.

A Linear SMPC Layer.

Keep traces of registered protocols. Attributes: registered_protocols (Dict[Any, Any]): Dictionary with the registered protocols.

The SGD Optimizer Class.

Metaclass for the MPCTensor and ShareTensors.