Quickstart PyTorch Lightning

In this federated learning tutorial we will learn how to train an AutoEncoder model on MNIST using Flower and PyTorch Lightning. It is recommended to create a virtual environment and run everything within a virtualenv.

Then, clone the code example directly from GitHub:

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

This will create a new directory called quickstart-pytorch-lightning containing the following files:

quickstart-pytorch-lightning
├── pytorchlightning_example
│   ├── client_app.py   # Defines your ClientApp
│   ├── server_app.py   # Defines your ServerApp
│   └── task.py         # Defines your model, training and data loading
├── pyproject.toml      # Project metadata like dependencies and configs
└── README.md

Next, activate your environment, then run:

# Navigate to the example directory
$ cd path/to/quickstart-pytorch-lightning

# Install project and dependencies
$ pip install -e .

By default, this project uses a local simulation profile that flwr run submits to a managed local SuperLink, which then executes the run with the Flower Simulation Runtime. It creates a federation of 4 nodes using FedAvg as the aggregation strategy. The dataset will be partitioned using Flower Dataset’s IidPartitioner. To run the project, do:

# Run with default arguments and stream logs
$ flwr run . --stream

Plain flwr run . submits the run, prints the run ID, and returns without streaming logs. For the full local workflow, see Run Flower Locally with a Managed SuperLink.

With default arguments you will see streamed output like this:

Successfully built flwrlabs.quickstart-pytorch-lightning.1-0-0.014c8eb3.fab
Starting local SuperLink on 127.0.0.1:39093...
Successfully started run 1859953118041441032
INFO :      Starting FedAvg strategy:
INFO :          ├── Number of rounds: 3
INFO :      [ROUND 1/3]
INFO :      configure_train: Sampled 2 nodes (out of 4)
INFO :      aggregate_train: Received 2 results and 0 failures
INFO :          └──> Aggregated MetricRecord: {'train_loss': 0.0487}
INFO :      configure_evaluate: Sampled 2 nodes (out of 4)
INFO :      aggregate_evaluate: Received 2 results and 0 failures
INFO :          └──> Aggregated MetricRecord: {'eval_loss': 0.0495}
INFO :      [ROUND 2/3]
INFO :      ...
INFO :      [ROUND 3/3]
INFO :      ...
INFO :      Strategy execution finished in 159.24s
INFO :      Final results:
INFO :          ServerApp-side Evaluate Metrics:
INFO :          {}

Each simulated ClientApp (two per round) will also log a summary of their local training process. Expect this output to be similar to:

# The left part indicates the process ID running the `ClientApp`
(ClientAppActor pid=38155) ┏━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
(ClientAppActor pid=38155) ┃        Test metric        ┃       DataLoader 0        ┃
(ClientAppActor pid=38155) ┡━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
(ClientAppActor pid=38155) │         test_loss         │   0.045175597071647644    │
(ClientAppActor pid=38155) └───────────────────────────┴───────────────────────────┘

You can also override the parameters defined in the [tool.flwr.app.config] section in pyproject.toml like this:

# Override some arguments
$ flwr run . --run-config num-server-rounds=5

팁

Check the 시뮬레이션 실행 documentation to learn more about how to configure and run Flower simulations.

참고

Check the source code of this tutorial in examples/quickstart-pytorch-lightning in the Flower GitHub repository.