Go through the tutorial as a group

[michellewang]

https://flower.ai/docs/framework/tutorial-quickstart-scikitlearn.html

• [x] load the MNIST dataset using Flower Datasets
• [x] Split the data in to train/test sets
• [x] Define logistic regression model and how it integrates with Flower interface
  • Define MnistClient which inherits from NumpyClient
  • NumPyClient makes it easier to implement the Client interface using using scikit-learn
• [x] Run server that expects 2 clients
• [x] Run 2 clients

Flower Server

• The number of federated learning rounds is set in fit_round() and the evaluation is defined in get_evaluate_fn(). The evaluation function is called after each federated learning round and gives you information about loss and accuracy.

[emmajanecorley]

Including some notes:

The server coordinates the training process. It sends the initial model parameters to the clients. It collects updated model parameters from clients after they have trained on their local data. It aggregates these updates to improve the global model. It repeats this process for a set number of rounds.

Next is create a client (client.py).

Each client has its own local data and model. It receives initial model parameters from the server. It trains the model on its local data. It sends the updated model parameters back to the server. It also evaluates the model using local test data and reports the results to the server.