U
- data type for sparseValues()
outputpublic final class RaggedTensorToSparse<U> extends PrimitiveOp
input=ragged.from_nested_row_splits(rt_dense_values, rt_nested_splits) output=SparseTensor(indices=sparse_indices, values=sparse_values, dense_shape=sparse_dense_shape)
operation
Modifier and Type | Method and Description |
---|---|
static <U,T extends Number> |
create(Scope scope,
Iterable<Operand<T>> rtNestedSplits,
Operand<U> rtDenseValues)
Factory method to create a class wrapping a new RaggedTensorToSparse operation.
|
Output<Long> |
sparseDenseShape()
`sparse_dense_shape` is a tight bounding box of the input `RaggedTensor`.
|
Output<Long> |
sparseIndices()
The indices for the `SparseTensor`.
|
Output<U> |
sparseValues()
The values of the `SparseTensor`.
|
equals, hashCode, op, toString
public static <U,T extends Number> RaggedTensorToSparse<U> create(Scope scope, Iterable<Operand<T>> rtNestedSplits, Operand<U> rtDenseValues)
scope
- current scopertNestedSplits
- The `row_splits` for the `RaggedTensor`.rtDenseValues
- The `flat_values` for the `RaggedTensor`.Copyright © 2022. All rights reserved.