Warning: This API is deprecated and will be removed in a future version of TensorFlow after the replacement is stable.

RaggedCross

public final class RaggedCross

Generates a feature cross from a list of tensors, and returns it as a RaggedTensor. See `tf.ragged.cross` for more details.

Public Methods

static <T, U extends Number> RaggedCross <T, U>
create ( Scope scope, Iterable< Operand <?>> raggedValues, Iterable< Operand <?>> raggedRowSplits, Iterable< Operand <Long>> sparseIndices, Iterable< Operand <?>> sparseValues, Iterable< Operand <Long>> sparseShape, Iterable< Operand <?>> denseInputs, String inputOrder, Boolean hashedOutput, Long numBuckets, Long hashKey, Class<T> outValuesType, Class<U> outRowSplitsType)
Factory method to create a class wrapping a new RaggedCross operation.
Output <U>
outputRowSplits ()
The `row_splits` for the returned `RaggedTensor`.
Output <T>
outputValues ()
The `values` for the returned `RaggedTensor`.

Inherited Methods

Public Methods

public static RaggedCross <T, U> create ( Scope scope, Iterable< Operand <?>> raggedValues, Iterable< Operand <?>> raggedRowSplits, Iterable< Operand <Long>> sparseIndices, Iterable< Operand <?>> sparseValues, Iterable< Operand <Long>> sparseShape, Iterable< Operand <?>> denseInputs, String inputOrder, Boolean hashedOutput, Long numBuckets, Long hashKey, Class<T> outValuesType, Class<U> outRowSplitsType)

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

Parameters
scope current scope
raggedValues The values tensor for each RaggedTensor input.
raggedRowSplits The row_splits tensor for each RaggedTensor input.
sparseIndices The indices tensor for each SparseTensor input.
sparseValues The values tensor for each SparseTensor input.
sparseShape The dense_shape tensor for each SparseTensor input.
denseInputs The tf.Tensor inputs.
inputOrder String specifying the tensor type for each input. The `i`th character in this string specifies the type of the `i`th input, and is one of: 'R' (ragged), 'D' (dense), or 'S' (sparse). This attr is used to ensure that the crossed values are combined in the order of the inputs from the call to tf.ragged.cross.
Returns
  • a new instance of RaggedCross

public Output <U> outputRowSplits ()

The `row_splits` for the returned `RaggedTensor`.

public Output <T> outputValues ()

The `values` for the returned `RaggedTensor`.