T
- data type for outputNestedSplits()
outputU
- data type for outputDenseValues()
outputpublic final class RaggedGather<T extends Number,U> extends PrimitiveOp
Outputs a `RaggedTensor` output composed from `output_dense_values` and `output_nested_splits`, such that:
output.shape = indices.shape + params.shape[1:]
output.ragged_rank = indices.shape.ndims + params.ragged_rank
output[i...j, d0...dn] = params[indices[i...j], d0...dn]
where
operation
Modifier and Type | Method and Description |
---|---|
static <T extends Number,U,V extends Number> |
create(Scope scope,
Iterable<Operand<T>> paramsNestedSplits,
Operand<U> paramsDenseValues,
Operand<V> indices,
Long OUTPUTRAGGEDRANK)
Factory method to create a class wrapping a new RaggedGather operation.
|
Output<U> |
outputDenseValues()
The `flat_values` for the returned RaggedTensor.
|
List<Output<T>> |
outputNestedSplits()
The `nested_row_splits` tensors that define the row-partitioning for the
returned RaggedTensor.
|
equals, hashCode, op, toString
public static <T extends Number,U,V extends Number> RaggedGather<T,U> create(Scope scope, Iterable<Operand<T>> paramsNestedSplits, Operand<U> paramsDenseValues, Operand<V> indices, Long OUTPUTRAGGEDRANK)
scope
- current scopeparamsNestedSplits
- The `nested_row_splits` tensors that define the row-partitioning for the
`params` RaggedTensor input.paramsDenseValues
- The `flat_values` for the `params` RaggedTensor. There was a terminology change
at the python level from dense_values to flat_values, so dense_values is the
deprecated name.indices
- Indices in the outermost dimension of `params` of the values that should be
gathered.OUTPUTRAGGEDRANK
- The ragged rank of the output RaggedTensor. `output_nested_splits` will contain
this number of `row_splits` tensors. This value should equal
`indices.shape.ndims + params.ragged_rank - 1`.public List<Output<T>> outputNestedSplits()
Copyright © 2022. All rights reserved.