client module

class client.Client(client_id, group=None, train_data={'x': [], 'y': []}, eval_data={'x': [], 'y': []}, model=None)

Bases: object

model

Returns this client reference to model being trained

num_test_samples

Number of test samples for this client.

Returns:Number of test samples for this client Return type:int

test(model)

Tests self.model on self.eval_data.

Returns:dict of metrics returned by the model.

train(num_epochs=1, batch_size=10, minibatch=None)

Trains on self.model using the client’s train_data.

Parameters:

• num_epochs – Number of epochs to train. Unsupported if minibatch is provided (minibatch has only 1 epoch)
• batch_size – Size of training batches.
• minibatch – fraction of client’s data to apply minibatch sgd, None to use FedAvg

Returns:

number of FLOPs executed in training process num_samples: number of samples used in training update: set of weights update_size: number of bytes in update

Return type:

comp