NVIDIA FLARE (NVIDIA Federated Learning Application Runtime Environment) is a domain-agnostic, open-source, extensible SDK that allows researchers and data scientists to adaptexisting ML/DL workflows (PyTorch, RAPIDS, Nemo, TensorFlow) to a federated paradigm; and enables platform developers to build a secure, privacy preserving offering for a distributed multi-party collaboration.

Federated learning allows multiple clients, each with their own data, to collaborate without sharing data. Different parties or institutions located throughtout the world can perform a set of tasks on their own local data, coordinated by a secure, central Federated Learning server, to build a global model. NVIDIA FLARE enables this collaborative workflow without ever needing to give external access to a participants’ local data.

NVIDIA FLARE is built on a componentized architecture that allows researchers to customize workflows to their liking and experiment with different ideas quickly.

With NVIDIA FLARE 2.1.0, High Availability (HA) and Multi-Job Execution introduce new concepts and change the way the system needs to be configured and operated. See conversion from 2.0 for details.