Federated Learning with Pandas and Flower (Quickstart Example)

[!CAUTION] This example uses Flower’s low-level API which remains a preview feature and subject to change. Both ClientApp and ServerApp operate directly on Message and RecordSet objects.

This introductory example to Flower uses Pandas, but deep knowledge of Pandas is not necessarily required to run the example. However, it will help you understand how to adapt Flower to your use case. This example uses Flower Datasets to download, partition and preprocess the Iris dataset. Running this example in itself is quite easy.

This example implements a form of Federated Analyics by which instead of training a model using locally available data, the nodes run a query on the data they own. In this example the query is to compute the histogram on specific columns of the dataset. These metrics are sent to the ServerApp for aggregation.

Set up the project

Clone the project

Start by cloning the example project.

git clone --depth=1 https://github.com/adap/flower.git _tmp \
		&& mv _tmp/examples/quickstart-pandas . \
		&& rm -rf _tmp && cd quickstart-pandas

This will create a new directory called quickstart-pandas with the following structure:

quickstart-pandas
├── pandas_example
│   ├── __init__.py
│   ├── client_app.py   # Defines your ClientApp
│   └── server_app.py   # Defines your ServerApp
├── pyproject.toml      # Project metadata like dependencies and configs
└── README.md

Install dependencies and project

Install the dependencies defined in pyproject.toml as well as the pandas_example package.

pip install -e .

Run the project

You can run your Flower project in both simulation and deployment mode without making changes to the code. If you are starting with Flower, we recommend you using the simulation mode as it requires fewer components to be launched manually. By default, flwr run will make use of the Simulation Engine.

Run with the Simulation Engine

flwr run .

You can also override some of the settings for your ClientApp and ServerApp defined in pyproject.toml. For example

flwr run . --run-config num-server-rounds=5

[!TIP] For a more detailed walk-through check our quickstart PyTorch tutorial

Run with the Deployment Engine

[!NOTE] An update to this example will show how to run this Flower application with the Deployment Engine and TLS certificates, or with Docker.