tff.learning.build_federated_averaging_process

Builds an iterative process that performs federated averaging.

tff.learning.build_federated_averaging_process(
    model_fn: Callable[[], tff.learning.Model],
    client_optimizer_fn: Callable[[], tf.keras.optimizers.Optimizer],
    server_optimizer_fn: Callable[[], tf.keras.optimizers.Optimizer] = DEFAULT_SERVER_OPTIMIZER_FN,
    client_weight_fn: Callable[[Any], tf.Tensor] = None,
    *,
    broadcast_process: Optional[tff.templates.MeasuredProcess] = None,
    aggregation_process: Optional[tff.templates.MeasuredProcess] = None,
    model_update_aggregation_factory: Optional[tff.aggregators.AggregationProcessFactory] = None,
    use_experimental_simulation_loop: bool = False
) -> tff.templates.IterativeProcess

Used in the notebooks

Used in the tutorials

This function creates a tff.templates.IterativeProcess that performs federated averaging on client models. The iterative process has the following methods:

  • initialize: A tff.Computation with the functional type signature ( -> S@SERVER), where S is a tff.learning.framework.ServerState representing the initial state of the server.
  • next: A tff.Computation with the functional type signature (<S@SERVER, {B*}@CLIENTS> -> <S@SERVER, T@SERVER>) where S is a tff.learning.framework.ServerState whose type matches that of the output of initialize, and {B*}@CLIENTS represents the client datasets, where B is the type of a single batch. This computation returns a tff.learning.framework.ServerState representing the updated server state and metrics that are the result of tff.learning.Model.federated_output_computation during client training and any other metrics from broadcast and aggregation processes.

Each time the next method is called, the server model is broadcast to each client using a broadcast function. For each client, one epoch of local training is performed via the tf.keras.optimizers.Optimizer.apply_gradients method of the client optimizer. Each client computes the difference between the client model after training and the initial broadcast model. These model deltas are then aggregated at the server using some aggregation function. The aggregate model delta is applied at the server by using the tf.keras.optimizers.Optimizer.apply_gradients method of the server optimizer.

model_fn A no-arg function that returns a tff.learning.Model. This method must not capture TensorFlow tensors or variables and use them. The model must be constructed entirely from scratch on each invocation, returning the same pre-constructed model each call will result in an error.
client_optimizer_fn A no-arg callable that returns a tf.keras.Optimizer.
server_optimizer_fn A no-arg callable that returns a tf.keras.Optimizer. By default, this uses tf.keras.optimizers.SGD with a learning rate of 1.0.
client_weight_fn Optional function that takes the output of model.report_local_outputs and returns a tensor providing the weight in the federated average of model deltas. If not provided, the default is the total number of examples processed on device.
broadcast_process a tff.templates.MeasuredProcess that broadcasts the model weights on the server to the clients. It must support the signature (input_values@SERVER -> output_values@CLIENT).
aggregation_process a tff.templates.MeasuredProcess that aggregates the model updates on the clients back to the server. It must support the signature ({input_values}@CLIENTS-> output_values@SERVER). Must be None if model_update_aggregation_factory is not None.
model_update_aggregation_factory An optional tff.aggregators.AggregationProcessFactory that constructs tff.templates.AggregationProcess for aggregating the client model updates on the server. If None, uses a default constructed tff.aggregators.MeanFactory, creating a stateless mean aggregation. Must be None if aggregation_process is not None.
use_experimental_simulation_loop Controls the reduce loop function for input dataset. An experimental reduce loop is used for simulation.

A tff.templates.IterativeProcess.