PartitionedOutput

public final class PartitionedOutput

An op that demultiplexes a tensor to be sharded by XLA to a list of partitioned

outputs outside the XLA computation.

Nested Classes

class PartitionedOutput.Options Optional attributes for PartitionedOutput  

Constants

String OP_NAME The name of this op, as known by TensorFlow core engine

Public Methods

static <T extends TType> PartitionedOutput<T>
create(Scope scope, Operand<T> inputs, Long numSplits, Options... options)
Factory method to create a class wrapping a new PartitionedOutput operation.
Iterator<Operand<T>>
List<Output<T>>
output()
A list of partitioned inputs which must have the same shape.
static PartitionedOutput.Options
partitionDim(Long partitionDim)

Inherited Methods

Constants

public static final String OP_NAME

The name of this op, as known by TensorFlow core engine

Constant Value: "TPUPartitionedOutput"

Public Methods

public static PartitionedOutput<T> create (Scope scope, Operand<T> inputs, Long numSplits, Options... options)

Factory method to create a class wrapping a new PartitionedOutput operation.

Parameters
scope current scope
inputs A tensor which represents the full shape of partitioned tensors.
options carries optional attributes values
Returns
  • a new instance of PartitionedOutput

public Iterator<Operand<T>> iterator ()

public List<Output<T>> output ()

A list of partitioned inputs which must have the same shape.

public static PartitionedOutput.Options partitionDim (Long partitionDim)

Parameters
partitionDim An integer describles which dimension is partitioned.