TPURoundRobin

public final class TPURoundRobin

Round-robin load balancing on TPU cores.

A load balancing op that round-robins among TPU cores.

This op round-robins between the integers in [0, NumTPUCoresVisiblePerHost]. It is useful for interfacing with TensorFlow ops that take as input a TPU core on which to execute computations, such as `TPUPartitionedCall`.

device_ordinal: An integer in [0, NumTPUCoresVisiblePerHost].

Public Methods

Output<Integer>
asOutput()
Returns the symbolic handle of a tensor.
static TPURoundRobin
create(Scope scope)
Factory method to create a class wrapping a new TPURoundRobin operation.
Output<Integer>
deviceOrdinal()

Inherited Methods

Public Methods

public Output<Integer> asOutput ()

Returns the symbolic handle of a tensor.

Inputs to TensorFlow operations are outputs of another TensorFlow operation. This method is used to obtain a symbolic handle that represents the computation of the input.

public static TPURoundRobin create (Scope scope)

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

Parameters
scope current scope
Returns
  • a new instance of TPURoundRobin

public Output<Integer> deviceOrdinal ()