tff.simulation.build_uniform_sampling_fn

Stay organized with collections Save and categorize content based on your preferences.

Builds the function for sampling from the input iterator at each round.

tff.simulation.build_uniform_sampling_fn(
    sample_range: Sequence[T],
    replace: bool = False,
    random_seed: Optional[int] = None
) -> Callable[[int, int], list[T]]

If an integer random_seed is provided, we set a random seed before sampling clients according to a multiplicative linear congruential generator (aka Lehmer generator, see 'The Art of Computer Programming, Vol. 3' by Donald Knuth for reference). This does not affect model initialization, shuffling, or other such aspects of the federated training process.

sample_range A 1-D array-like sequence, to be used as input to np.random.choice. Samples are generated randomly from the elements of the sequence.
replace A boolean indicating whether the sampling is done with replacement (True) or without replacement (False).
random_seed If an integer, it is used as a random seed for the client sampling process. If None, a nondeterministic seed is used.

A function that takes as input an integer round_num and integer size and returns a list of size elements sampled (pseudo-)randomly from the input sample_range.