InfeedEnqueueTuple

public final class InfeedEnqueueTuple

Feeds multiple Tensor values into the computation as an XLA tuple.

Nested Classes

class InfeedEnqueueTuple.Options Optional attributes for InfeedEnqueueTuple  

Public Methods

static InfeedEnqueueTuple
create(Scope scope, Iterable<Operand<?>> inputs, List<Shape> shapes, Options... options)
Factory method to create a class wrapping a new InfeedEnqueueTuple operation.
static InfeedEnqueueTuple.Options
deviceOrdinal(Long deviceOrdinal)
static InfeedEnqueueTuple.Options
layouts(List<Long> layouts)

Inherited Methods

Public Methods

public static InfeedEnqueueTuple create (Scope scope, Iterable<Operand<?>> inputs, List<Shape> shapes, Options... options)

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

Parameters
scope current scope
inputs A list of tensors that will be provided using the infeed mechanism.
shapes The shapes of each tensor in `inputs`.
options carries optional attributes values
Returns
  • a new instance of InfeedEnqueueTuple

public static InfeedEnqueueTuple.Options deviceOrdinal (Long deviceOrdinal)

Parameters
deviceOrdinal The TPU device to use. This should be -1 when the Op is running on a TPU device, and >= 0 when the Op is running on the CPU device.

public static InfeedEnqueueTuple.Options layouts (List<Long> layouts)

Parameters
layouts A vector holding the requested layout in minor-to-major sequence for all the tuple shapes, in the order the shapes appear in the "shapes" input. The layout elements for a sub-shape can be set to -1, in which case the corresponding layout will be computed by the infeed operation.