Builds a new learning process which constructs datasets on clients.
tff.simulation.compose_dataset_computation_with_learning_process(
dataset_computation: tff.Computation,
process: tff.learning.templates.LearningProcess
) -> tff.learning.templates.LearningProcess
This functionality is identical to
tff.simulation.compose_dataset_computation_with_iterative_process, except
that all public attributes of the process (except for initialize and next)
are also preserved (eg. LearningProcess.get_model_weights).
Args |
|---|
dataset_computation
|
An instance of tff.Computation which accepts some
parameter and returns an element of tff.SequenceType.
|
process
|
An instance of tff.learning.templates.LearningProcess whose
next function accepts exactly one federated dataset (ie. something of
type {B*}@CLIENTS, where B is equivalent to the return type of
dataset_computation).
|
View source on GitHub