federated

Manages the federated communication and training of models.

Federated algorithm plugins can also be imported from this package.

Module

Submodules

• bitfount.federated.aggregators - Model parameter aggregators for Federated Averaging.
• bitfount.federated.algorithms - Algorithms for remote processing of data.
• bitfount.federated.authorisation_checkers - Authorisation Checkers.
• bitfount.federated.early_stopping - Federated early stopping.
• bitfount.federated.encryption - Symmetric and asymmetric encryption functions.
• bitfount.federated.exceptions - Custom exceptions for the federated package.
• bitfount.federated.helper - Helper functions.
• bitfount.federated.keys_setup - Module for setting up the pod keys.
• bitfount.federated.logging - Federated logging.
• bitfount.federated.mixins - MixIn classes for compatible models with the federated algorithms.
• bitfount.federated.model_reference - References to custom models.
• bitfount.federated.modeller - Modeller for dispatching tasks.
• bitfount.federated.monitoring - Contains functionality for interacting with the task monitoring service.
• bitfount.federated.pod - Pods for responding to tasks.
• bitfount.federated.pod_db_utils - Utilities for the Pod results database.
• bitfount.federated.pod_response_message - Pod Response Message.
• bitfount.federated.pod_vitals - Module for creating Pod Vitals webserver.
• bitfount.federated.privacy - Modules for federated privacy approaches.
• bitfount.federated.protocols - Protocols for inter-machine communication.
• bitfount.federated.roles - Contains mixin for role-finding code.
• bitfount.federated.secure - Secure aggregation.
• bitfount.federated.shim - Backend-agnostisc shims.
• bitfount.federated.task_requests - Dataclasses and functionality for task request details/messages.
• bitfount.federated.transport - Transport layer classes for communicating tasks between pods and modellers.
• bitfount.federated.types - Useful types for Federated Learning.
• bitfount.federated.utils - Useful objects/functions for federated learning.
• bitfount.federated.worker - Workers for handling task running on pods.

---