ML Community Day is November 9! Join us for updates from TensorFlow, JAX, and more Learn more

Builds the tff.Computation for heavy-hitters discovery with IBLT.

capacity The capacity of the IBLT sketch. Defaults to 1000.
max_string_length The maximum length of a string in the IBLT. Defaults to 10. Must be positive.
repetitions The number of repetitions in IBLT data structure (must be >= 3). Defaults to 3. Must be at least 3.
seed An integer seed for hash functions. Defaults to 0.
dtype A tensorflow data type which determines the type of the IBLT values. Must be tf.int32 or tf.int64. Defaults to tf.int64.
max_heavy_hitters The maximum number of items to return. If the decoded results have more than this number of items, will order decreasingly by the estimated counts and return the top max_heavy_hitters items. Default max_heavy_hitters == None, which means to return all the heavy hitters in the result.
max_words_per_user The maximum number of words each client is allowed to contribute. If not None, must be a positive integer. Defaults to None, which means all the clients contribute all their words.
k_anonymity Only return words that appear in at least k clients. Must be a positive integer. Defaults to 1.
secure_sum_bitwidth The bitwidth used for secure sum. The default value is None, which disables secure sum. If not None, must be in the range [1,62]. See tff.federated_secure_sum_bitwidth.
batch_size The number of elements in each batch of the dataset. Defaults to 1, means the input dataset is processed by Must be a positive.
multi_contribution Whether each client is allowed to contribute multiple counts or only a count of one for each unique word. Defaults to True.
decode_iblt_fn A function to decode key-value pairs from an IBLT sketch. Defaults to None, in this case decode_iblt_fn will be set to iblt.decode_iblt_tf.

A tff.Computation that performs federated heavy hitter discovery.

ValueError if parameters don't meet expectations.