Quickstart scikit-learn¶
In this federated learning tutorial we will learn how to train a Logistic Regression on MNIST using Flower and scikit-learn. It is recommended to create a virtual environment and run everything within a virtualenv.
Let’s use flwr new to create a complete Flower+scikit-learn project. It will
generate all the files needed to run, by default with the Flower Simulation Engine, a
federation of 10 nodes using FedAvg The dataset will be partitioned using
Flower Datasets’s IidPartitioner
Now that we have a rough idea of what this example is about, let’s get started. First, install Flower in your new environment:
# In a new Python environment
$ pip install flwr
Then, run the command below. You will be prompted to select one of the available
templates (choose sklearn), give a name to your project, and type in your developer
name:
$ flwr new
After running it you’ll notice a new directory with your project name has been created. It should have the following structure:
<your-project-name>
├── <your-project-name>
│ ├── __init__.py
│ ├── 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
If you haven’t yet installed the project and its dependencies, you can do so by:
# From the directory where your pyproject.toml is
$ pip install -e .
To run the project, do:
# Run with default arguments
$ flwr run .
With default arguments you will see an output like this one:
Loading project configuration...
Success
INFO : Starting FedAvg strategy:
INFO : ├── Number of rounds: 3
INFO : ├── ArrayRecord (0.06 MB)
INFO : ├── ConfigRecord (train): (empty!)
INFO : ├── ConfigRecord (evaluate): (empty!)
INFO : ├──> Sampling:
INFO : │ ├──Fraction: train (1.00) | evaluate ( 1.00)
INFO : │ ├──Minimum nodes: train (2) | evaluate (2)
INFO : │ └──Minimum available nodes: 2
INFO : └──> Keys in records:
INFO : ├── Weighted by: 'num-examples'
INFO : ├── ArrayRecord key: 'arrays'
INFO : └── ConfigRecord key: 'config'
INFO :
INFO :
INFO : [ROUND 1/3]
INFO : configure_train: Sampled 10 nodes (out of 10)
INFO : aggregate_train: Received 10 results and 0 failures
INFO : └──> Aggregated MetricRecord: {'train_logloss': 1.3937176081476854}
INFO : configure_evaluate: Sampled 10 nodes (out of 10)
INFO : aggregate_evaluate: Received 10 results and 0 failures
INFO : └──> Aggregated MetricRecord: {'test_logloss': 1.23306, 'accuracy': 0.69154, 'precision': 0.68659, 'recall': 0.68046, 'f1': 0.65752}
INFO :
INFO : [ROUND 2/3]
INFO : configure_train: Sampled 10 nodes (out of 10)
INFO : aggregate_train: Received 10 results and 0 failures
INFO : └──> Aggregated MetricRecord: {'train_logloss': 0.8565170774432291}
INFO : configure_evaluate: Sampled 10 nodes (out of 10)
INFO : aggregate_evaluate: Received 10 results and 0 failures
INFO : └──> Aggregated MetricRecord: {'test_logloss': 0.8805, 'accuracy': 0.73425, 'precision': 0.792371, 'recall': 0.7329, 'f1': 0.70438}
INFO :
INFO : [ROUND 3/3]
INFO : configure_train: Sampled 10 nodes (out of 10)
INFO : aggregate_train: Received 10 results and 0 failures
INFO : └──> Aggregated MetricRecord: {'train_logloss': 0.703260769576}
INFO : configure_evaluate: Sampled 10 nodes (out of 10)
INFO : aggregate_evaluate: Received 10 results and 0 failures
INFO : └──> Aggregated MetricRecord: {'test_logloss': 0.70207, 'accuracy': 0.77250, 'precision': 0.82201, 'recall': 0.76348, 'f1': 0.75069}
INFO :
INFO : Strategy execution finished in 17.87s
INFO :
INFO : Final results:
INFO :
INFO : Global Arrays:
INFO : ArrayRecord (0.060 MB)
INFO :
INFO : Aggregated ClientApp-side Train Metrics:
INFO : { 1: {'train_logloss': '1.3937e+00'},
INFO : 2: {'train_logloss': '8.5652e-01'},
INFO : 3: {'train_logloss': '7.0326e-01'}}
INFO :
INFO : Aggregated ClientApp-side Evaluate Metrics:
INFO : { 1: { 'accuracy': '6.9158e-01',
INFO : 'f1': '6.5752e-01',
INFO : 'precision': '6.8659e-01',
INFO : 'recall': '6.8046e-01',
INFO : 'test_logloss': '1.2331e+00'},
INFO : 2: { 'accuracy': '7.3425e-01',
INFO : 'f1': '7.0439e-01',
INFO : 'precision': '7.9237e-01',
INFO : 'recall': '7.3295e-01',
INFO : 'test_logloss': '8.8056e-01'},
INFO : 3: { 'accuracy': '7.7250e-01',
INFO : 'f1': '7.5069e-01',
INFO : 'precision': '8.2201e-01',
INFO : 'recall': '7.6348e-01',
INFO : 'test_logloss': '7.0208e-01'}}
INFO :
INFO : ServerApp-side Evaluate Metrics:
INFO : {}
INFO :
Saving final model to disk...
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 local-epochs=2"
What follows is an explanation of each component in the project you just created:
dataset partition, the model, defining the ClientApp and defining the ServerApp.
The Data¶
This tutorial uses Flower Datasets to easily download and partition the MNIST dataset. In this example you’ll make
use of the IidPartitioner to generate num_partitions partitions. You can choose
other partitioners available in Flower Datasets. Each ClientApp will call this
function to create dataloaders with the data that correspond to their data partition.
partitioner = IidPartitioner(num_partitions=num_partitions)
fds = FederatedDataset(
dataset="mnist",
partitioners={"train": partitioner},
)
dataset = fds.load_partition(partition_id, "train").with_format("numpy")
X, y = dataset["image"].reshape((len(dataset), -1)), dataset["label"]
# Split the on edge data: 80% train, 20% test
X_train, X_test = X[: int(0.8 * len(X))], X[int(0.8 * len(X)) :]
y_train, y_test = y[: int(0.8 * len(y))], y[int(0.8 * len(y)) :]
The Model¶
We define the LogisticRegression model from scikit-learn in the get_model()
function:
def get_model(penalty: str, local_epochs: int):
return LogisticRegression(
penalty=penalty,
max_iter=local_epochs,
warm_start=True,
)
The ClientApp¶
The main changes we have to make to use Scikit-learn with Flower have to do with
converting the ArrayRecord received in the Message into numpy ndarrays
and then use them to set the model parameters. After training, another auxiliary
function can be used to extract then pack the updated numpy ndarrays into a Message
from the ClientApp. We can make use of built-in methods in the ArrayRecord to make
these conversions:
@app.train()
def train(msg: Message, context: Context):
# Load the model
model = get_model() # construct your scikit-learn model
# Extract the ArrayRecord from Message and convert to numpy ndarrays
ndarrays = msg.content["arrays"].to_numpy_ndarrays()
# Set the model parameters with auxhiliary function
set_model_params(model, ndarrays)
# Train the model
...
# Extract the updated model parameters with auxhiliary function
updated_ndarrays = get_model_params(model)
# Pack the updated parameters into an ArrayRecord
model_record = ArrayRecord(updated_ndarrays)
The rest of the functionality is directly inspired by the centralized case. The
ClientApp comes with three core methods (train, evaluate, and query)
that we can implement for different purposes. For example: train to train the
received model using the local data; evaluate to assess its performance of the
received model on a validation set; and query to retrieve information about the node
executing the ClientApp. In this tutorial we will only make use of train and
evaluate.
Let’s see how the train method can be implemented. It receives as input arguments a
Message from the ServerApp. By default it carries:
an
ArrayRecordwith the arrays of the model to federate. By default they can be retrieved with key"arrays"when accessing the message content.a
ConfigRecordwith the configuration sent from theServerApp. By default it can be retrieved with key"config"when accessing the message content.
The train method also receives the Context, giving access to configs for your
run and node. The run config hyperparameters are defined in the pyproject.toml of
your Flower App. The node config can only be set when running Flower with the Deployment
Runtime and is not directly configurable during simulations.
app = ClientApp()
@app.train()
def train(msg: Message, context: Context):
"""Train the model on local data."""
# Create LogisticRegression Model
penalty = context.run_config["penalty"]
local_epochs = context.run_config["local-epochs"]
model = get_model(penalty, local_epochs)
# Setting initial parameters, akin to model.compile for keras models
set_initial_params(model)
# Apply received pararameters
ndarrays = msg.content["arrays"].to_numpy_ndarrays()
set_model_params(model, ndarrays)
# Load the data
partition_id = context.node_config["partition-id"]
num_partitions = context.node_config["num-partitions"]
X_train, _, y_train, _ = load_data(partition_id, num_partitions)
# Ignore convergence failure due to low local epochs
with warnings.catch_warnings():
warnings.simplefilter("ignore")
# Train the model on local data
model.fit(X_train, y_train)
# Let's compute train loss
y_train_pred_proba = model.predict_proba(X_train)
train_logloss = log_loss(y_train, y_train_pred_proba)
# Construct and return reply Message
ndarrays = get_model_params(model)
model_record = ArrayRecord(ndarrays)
metrics = {"num-examples": len(X_train), "train_logloss": train_logloss}
metric_record = MetricRecord(metrics)
content = RecordDict({"arrays": model_record, "metrics": metric_record})
return Message(content=content, reply_to=msg)
The @app.evaluate method mirrors train but only evaluates the received model on
the local validation set. It returns a MetricRecord containing the evaluation loss
and accuracy and does not include the model weights, since they are not modified during
evaluation.
The ServerApp¶
To construct a ServerApp we define its @app.main() method. This method
receive as input arguments:
a
Gridobject that will be used to interface with the nodes running theClientAppto involve them in a round of train/evaluate/query or other.a
Contextobject that provides access to the run configuration.
In this example we use the FedAvg and configure it with a specific value of
fraction_train which is read from the run config. You can find the default value
defined in the pyproject.toml. Then, the execution of the strategy is launched when
invoking its start method. To it we pass:
the
Gridobject.an
ArrayRecordcarrying a randomly initialized model that will serve as the global model to federated.a
ConfigRecordwith the training hyperparameters to be sent to the clients. The strategy will also insert the current round number in this config before sending it to the participating nodes.the
num_roundsparameter specifying how many rounds ofFedAvgto perform.
app = ServerApp()
@app.main()
def main(grid: Grid, context: Context) -> None:
"""Main entry point for the ServerApp."""
# Read run config
num_rounds: int = context.run_config["num-server-rounds"]
# Create LogisticRegression Model
penalty = context.run_config["penalty"]
local_epochs = context.run_config["local-epochs"]
model = get_model(penalty, local_epochs)
# Setting initial parameters, akin to model.compile for keras models
set_initial_params(model)
# Construct ArrayRecord representation
arrays = ArrayRecord(get_model_params(model))
# Initialize FedAvg strategy
strategy = FedAvg(fraction_train=1.0, fraction_evaluate=1.0)
# Start strategy, run FedAvg for `num_rounds`
result = strategy.start(
grid=grid,
initial_arrays=arrays,
num_rounds=num_rounds,
)
# Save final model parameters
print("\nSaving final model to disk...")
ndarrays = result.arrays.to_numpy_ndarrays()
set_model_params(model, ndarrays)
joblib.dump(model, "logreg_model.pkl")
Congratulations! You’ve successfully built and run your first federated learning system in scikit-learn on the MNIST dataset using the new Message API.
Note
Check the source code of another Flower App using scikit-learn in the Flower
GitHub repository.