DepthToSpace - version 11#

This page documents version 11 of operator DepthToSpace. See DepthToSpace for the latest version (since version 13).

  • Domain: ai.onnx

  • Since version: 11

DepthToSpace rearranges (permutes) 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. By default, mode = DCR. In the DCR mode, elements along the depth dimension from the input tensor are rearranged in the following order: depth, column, and then row. The output y is computed from the input x as below:

b, c, h, w = x.shape

tmp = np.reshape(x, [b, blocksize, blocksize, c // (blocksize**2), h, w])

tmp = np.transpose(tmp, [0, 3, 4, 1, 5, 2])

y = np.reshape(tmp, [b, c // (blocksize**2), h * blocksize, w * blocksize])

In the CRD mode, elements along the depth dimension from the input tensor are rearranged in the following order: column, row, and the depth. The output y is computed from the input x as below:

b, c, h, w = x.shape

tmp = np.reshape(x, [b, c // (blocksize ** 2), blocksize, blocksize, h, w])

tmp = np.transpose(tmp, [0, 1, 4, 2, 5, 3])

y = np.reshape(tmp, [b, c // (blocksize ** 2), h * blocksize, w * blocksize])

Inputs

  • input (T): Input tensor of [N,C,H,W], where N is the batch axis, C is the channel or depth, H is the height and W is the width.

Outputs

  • output (T): Output tensor of [N, C/(blocksize * blocksize), H * blocksize, W * blocksize].

Attributes

  • blocksize (int): Blocks of [blocksize, blocksize] are moved.

  • mode (string): DCR (default) for depth-column-row order re-arrangement. Use CRD for column-row-depth order.

Type Constraints

  • T: Constrain input and output types to all tensor types. Allowed types: tensor(bool), tensor(complex128), tensor(complex64), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8).

Differences with previous version (1)#

SchemaDiff: DepthToSpace (domain 'ai.onnx')

  • old version: 1

  • new version: 11

  • breaking: no

Attributes:

  • added ‘mode’: type=STRING; required=False; default=DCR

Documentation:

  • line similarity: 0.19 (+24/-1 lines)

--- DepthToSpace v1
+++ DepthToSpace v11
@@ -1,4 +1,27 @@
 DepthToSpace rearranges (permutes) 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.
+and width dimensions. By default, `mode` = `DCR`.
+In the DCR mode, elements along the depth dimension from the input tensor are rearranged in the
+following order: depth, column, and then row. The output y is computed from the input x as below:
+
+b, c, h, w = x.shape
+
+tmp = np.reshape(x, [b, blocksize, blocksize, c // (blocksize**2), h, w])
+
+tmp = np.transpose(tmp, [0, 3, 4, 1, 5, 2])
+
+y = np.reshape(tmp, [b, c // (blocksize**2), h * blocksize, w * blocksize])
+
+
+In the CRD mode, elements along the depth dimension from the input tensor are rearranged in the
+following order: column, row, and the depth. The output y is computed from the input x as below:
+
+b, c, h, w = x.shape
+
+tmp = np.reshape(x, [b, c // (blocksize ** 2), blocksize, blocksize, h, w])
+
+tmp = np.transpose(tmp, [0, 1, 4, 2, 5, 3])
+
+y = np.reshape(tmp, [b, c // (blocksize ** 2), h * blocksize, w * blocksize])
+