Skip to content

Flower Runtime

The Flower runtime enables you to execute federated learning tasks using the Flower framework. It registers Function kinds flower-app, flower-client, and flower-server and supports various actions for federated learning workflows including training, building, and deploying federated applications.

  • flower-app: Execute Flower simulations
  • flower-client: Define a Flower client for federated learning participation
  • flower-server: Define a Flower server to coordinate federated learning

Prerequisites

Supported Python versions:

  • Python ≥ 3.9, < 3.13

Required packages:

  • digitalhub-runtime-flower

Install from PyPI:

pip install digitalhub-runtime-flower

For local execution (simulation mode):

pip install digitalhub-runtime-flower[local]

Usage overview

To execute federated learning tasks on the platform:

  1. Implement your Flower application (client and server).
  2. Create a Function resource that references your Flower code.
  3. Call function.run() to execute the federated learning task.

See how to for detailed instructions on which kind of applications you can execute and how to implement and deploy your Flower application. See Examples for code samples.