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 simulationsflower-client: Define a Flower client for federated learning participationflower-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:
For local execution (simulation mode):
Usage overview
To execute federated learning tasks on the platform:
- Implement your Flower application (client and server).
- Create a
Functionresource that references your Flower code. - 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.