T
- data type for output()
output@Operator public final class BatchToSpaceNd<T> extends PrimitiveOp implements Operand<T>
This operation reshapes the "batch" dimension 0 into `M + 1` dimensions of shape `block_shape + [batch]`, interleaves these blocks back into the grid defined by the spatial dimensions `[1, ..., M]`, to obtain a result with the same rank as the input. The spatial dimensions of this intermediate result are then optionally cropped according to `crops` to produce the output. This is the reverse of SpaceToBatch. See below for a precise description.
operation
Modifier and Type | Method and Description |
---|---|
Output<T> |
asOutput()
Returns the symbolic handle of a tensor.
|
static <T,U extends Number,V extends Number> |
create(Scope scope,
Operand<T> input,
Operand<U> blockShape,
Operand<V> crops)
Factory method to create a class wrapping a new BatchToSpaceNd operation.
|
Output<T> |
output() |
equals, hashCode, op, toString
public static <T,U extends Number,V extends Number> BatchToSpaceNd<T> create(Scope scope, Operand<T> input, Operand<U> blockShape, Operand<V> crops)
scope
- current scopeinput
- N-D with shape `input_shape = [batch] + spatial_shape + remaining_shape`,
where spatial_shape has M dimensions.blockShape
- 1-D with shape `[M]`, all values must be >= 1.crops
- 2-D with shape `[M, 2]`, all values must be >= 0.
`crops[i] = [crop_start, crop_end]` specifies the amount to crop from input
dimension `i + 1`, which corresponds to spatial dimension `i`. It is
required that
`crop_start[i] + crop_end[i] <= block_shape[i] * input_shape[i + 1]`.
This operation is equivalent to the following steps:
1. Reshape `input` to `reshaped` of shape: [block_shape[0], ..., block_shape[M-1], batch / prod(block_shape), input_shape[1], ..., input_shape[N-1]]
2. Permute dimensions of `reshaped` to produce `permuted` of shape [batch / prod(block_shape),
input_shape[1], block_shape[0], ..., input_shape[M], block_shape[M-1],
input_shape[M+1], ..., input_shape[N-1]]
3. Reshape `permuted` to produce `reshaped_permuted` of shape [batch / prod(block_shape),
input_shape[1] * block_shape[0], ..., input_shape[M] * block_shape[M-1],
input_shape[M+1], ..., input_shape[N-1]]
4. Crop the start and end of dimensions `[1, ..., M]` of `reshaped_permuted` according to `crops` to produce the output of shape: [batch / prod(block_shape),
input_shape[1] * block_shape[0] - crops[0,0] - crops[0,1], ..., input_shape[M] * block_shape[M-1] - crops[M-1,0] - crops[M-1,1],
input_shape[M+1], ..., input_shape[N-1]]
Some examples:
(1) For the following input of shape `[4, 1, 1, 1]`, `block_shape = [2, 2]`, and `crops = [[0, 0], [0, 0]]`:
[[[[1]]], [[[2]]], [[[3]]], [[[4]]]]
The output tensor has shape `[1, 2, 2, 1]` and value:
x = [[[[1], [2]], [[3], [4]]]]
(2) For the following input of shape `[4, 1, 1, 3]`, `block_shape = [2, 2]`, and
`crops = [[0, 0], [0, 0]]`:
[[[[1, 2, 3]]], [[[4, 5, 6]]], [[[7, 8, 9]]], [[[10, 11, 12]]]]
The output tensor has shape `[1, 2, 2, 3]` and value:
x = [[[[1, 2, 3], [4, 5, 6]],
[[7, 8, 9], [10, 11, 12]]]]
(3) For the following input of shape `[4, 2, 2, 1]`, `block_shape = [2, 2]`, and
`crops = [[0, 0], [0, 0]]`:
x = [[[[1], [3]], [[9], [11]]],
[[[2], [4]], [[10], [12]]],
[[[5], [7]], [[13], [15]]],
[[[6], [8]], [[14], [16]]]]
The output tensor has shape `[1, 4, 4, 1]` and value:
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]],
[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
(4) For the following input of shape `[8, 1, 3, 1]`, `block_shape = [2, 2]`, and
`crops = [[0, 0], [2, 0]]`:
x = [[[[0], [1], [3]]], [[[0], [9], [11]]],
[[[0], [2], [4]]], [[[0], [10], [12]]],
[[[0], [5], [7]]], [[[0], [13], [15]]],
[[[0], [6], [8]]], [[[0], [14], [16]]]]
The output tensor has shape `[2, 2, 4, 1]` and value:
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]]],
[[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
public Output<T> asOutput()
Operand
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.
asOutput
in interface Operand<T>
OperationBuilder.addInput(Output)
Copyright © 2022. All rights reserved.