partitioner#

Flower Datasets Partitioner package.

Classes

IidPartitioner(num_partitions)

Partitioner creates each partition sampled randomly from the dataset.

Partitioner()

The base partitioner class that enables obtaining federated partitions.

NaturalIdPartitioner(partition_by)

Partitioner for a dataset that can be divided by a column with partition ids.

DirichletPartitioner(num_partitions, ...[, ...])

Partitioner based on Dirichlet distribution.

SizePartitioner(num_partitions, ...)

Base class for the deterministic size partitioning based on the partition_id.

LinearPartitioner(num_partitions)

Partitioner creates partitions of size that are linearly correlated with id.

InnerDirichletPartitioner(partition_sizes, ...)

Partitioner based on Dirichlet distribution.

SquarePartitioner(num_partitions)

Partitioner creates partitions of size that are correlated with squared id.

ShardPartitioner(num_partitions, partition_by)

Partitioner based on shard of (typically) unique classes.

ExponentialPartitioner(num_partitions)

Partitioner creates partitions of size that are correlated with exp(id).