tensorflow::ops::DepthToSpace
#include <array_ops.h>
DepthToSpace for tensors of type T.
Summary
Rearranges data from depth into blocks of spatial data. This is the reverse transformation of SpaceToDepth. More specifically, this op outputs a copy of the input tensor where values from the depth
dimension are moved in spatial blocks to the height
and width
dimensions. The attr block_size
indicates the input block size and how the data is moved.
 Chunks of data of size
block_size * block_size
from depth are rearranged into nonoverlapping blocks of sizeblock_size x block_size
 The width the output tensor is
input_depth * block_size
, whereas the height isinput_height * block_size
.  The depth of the input tensor must be divisible by
block_size * block_size
.
That is, assuming the input is in the shape: [batch, height, width, depth]
, the shape of the output will be: [batch, height*block_size, width*block_size, depth/(block_size*block_size)]
This operation requires that the input tensor be of rank 4, and that block_size
be >=1 and that block_size * block_size
be a divisor of the input depth.
This operation is useful for resizing the activations between convolutions (but keeping all data), e.g. instead of pooling. It is also useful for training purely convolutional models.
For example, given this input of shape [1, 1, 1, 4]
, and a block size of 2:
```prettyprint x = [[[[1, 2, 3, 4]]]]
```
This operation will output a tensor of shape [1, 2, 2, 1]
:
```prettyprint [[[[1], [2]], [[3], [4]]]] ```
Here, the input has a batch of 1 and each batch element has shape [1, 1, 4]
, the corresponding output will have 2x2 elements and will have a depth of 1 channel (1 = 4 / (block_size * block_size)
). The output element shape is [2, 2, 1]
.
For an input tensor with larger depth, here of shape [1, 1, 1, 12]
, e.g.
```prettyprint x = [[[[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]]]] ```
This operation, for block size of 2, will return the following tensor of shape [1, 2, 2, 3]
```prettyprint [[[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]]]
```
Similarly, for the following input of shape [1 2 2 4]
, and a block size of 2:
```prettyprint x = [[[[1, 2, 3, 4], [5, 6, 7, 8]], [[9, 10, 11, 12], [13, 14, 15, 16]]]] ```
the operator will return the following tensor of shape [1 4 4 1]
:
```prettyprint x = [[ [1], [2], [5], [6]], [ [3], [4], [7], [8]], [ [9], [10], [13], [14]], [ [11], [12], [15], [16]]]
```
Arguments:
 scope: A Scope object
 block_size: The size of the spatial block, same as in Space2Depth.
Returns:
Output
: The output tensor.
Constructors and Destructors 


DepthToSpace(const ::tensorflow::Scope & scope, ::tensorflow::Input input, int64 block_size)

Public attributes 


output

Public functions 


node() const

::tensorflow::Node *

operator::tensorflow::Input() const


operator::tensorflow::Output() const


Public attributes
output
::tensorflow::Output output
Public functions
DepthToSpace
DepthToSpace( const ::tensorflow::Scope & scope, ::tensorflow::Input input, int64 block_size )
node
::tensorflow::Node * node() const
operator::tensorflow::Input
operator::tensorflow::Input() const
operator::tensorflow::Output
operator::tensorflow::Output() const