💻 Using the CLI

P2PFL provides a command-line interface (CLI) to simplify running experiments. You can launch the CLI using python -m p2pfl. This interface provides a convenient way to explore and run different federated learning experiments without manually writing code for each node. You can easily switch to different communication protocols, aggregators, and ML frameworks.

⌨️ Main Commands

Command	Description	Availability
experiment	Run experiments on the p2pfl platform.	✅
launch	Launch a new node in the p2pfl network.	🔜 Coming Soon
login	Authenticate with the p2pfl platform using your API token.	🔜 Coming Soon
remote	Interact with a remote node in the p2pfl network.	🔜 Coming Soon

🧪 Example: experiment Command

This command allows you to interact with pre-built examples. Here’s how you can use it:

• List available examples: python -m p2pfl experiment list. This will display a table of available examples with their descriptions.

• Run an example: python -m p2pfl experiment run <example_name> [options].This will run the specified example.

For instance, to run the mnist example for 2 rounds of training with 1 epoch each:

python -m p2pfl experiment run mnist --rounds 2 --epochs 1

This command will start the mnist experiment with the specified options and the default components. To see the available options for the mnist experiment, you can run:

python -m p2pfl experiment help mnist