mindspore.mint.nn.functional.conv_transpose2d

View Source On Gitee
mindspore.mint.nn.functional.conv_transpose2d(input, weight, bias=None, stride=1, padding=0, output_padding=0, groups=1, dilation=1)[source]

Applies a 2D transposed convolution operator over an input image composed of several input planes, sometimes also called deconvolution (although it is not an actual deconvolution).

Refer to mindspore.mint.nn.ConvTranspose2d for more details.

Warning

  • This is an experimental API that is subject to change or deletion.

  • In the scenario where inputs are non-contiguous, output_padding must be less than stride .

  • For Atlas training products, when the dtype of input is float32, the groups only supports 1.

Parameters

  • input (Tensor) – input tensor of shape \((minibatch, in\_channels, iH, iW)\) or \((in\_channels, iH, iW)\) .

  • weight (Tensor) – filters of shape \((in\_channels, \frac{out\_channels}{\text{groups}}, kH, kW)\) .

  • bias (Tensor, optional) – bias of shape \((out\_channels)\) . Default: None .

  • stride (Union[int, tuple(int), list[int]], optional) – the stride of the convolving kernel. Can be a single number or a tuple \((sH, sW)\) . Default: 1 .

  • padding (Union[int, tuple(int), list[int]], optional) – \(dilation * (kernel\_size - 1) - padding\) zero-padding will be added to both sides of each dimension in the input. Can be a single number or a tuple \((padH, padW)\) . Default: 0 .

  • output_padding (Union[int, tuple(int), list[int]], optional) – additional size added to one side of each dimension in the output shape. Can be a single number or a tuple \((out\_padH, out\_padW)\) . The value of output_padding must be less than stride or dilation . Default: 0 .

  • groups (int, optional) – split input into groups, \(in\_channels\) should be divisible by the number of groups. Default: 1 .

  • dilation (Union[int, tuple(int), list[int]], optional) – the spacing between kernel elements. Can be a single number or a tuple \((dH, dW)\) . Default: 1 .

Returns

Tensor of shape \((minibatch, out\_channels, oH, oW)\) or \((out\_channels, oH, oW)\) , where

\[oH = (iH - 1) \times sH - 2 \times padH + dH \times (kH - 1) + out\_padH + 1\]
\[oW = (iW - 1) \times sW - 2 \times padW + dW \times (kW - 1) + out\_padW + 1\]

Raises

  • TypeError – If stride, padding, output_padding or dilation is neither an int nor a tuple or a list.

  • TypeError – If groups is not an int.

  • ValueError – If the shape of bias is not \((out\_channels)\) .

  • ValueError – If stride or dilation is less than 1.

  • ValueError – If padding or output_padding is less than 0.

  • ValueError – If stride, padding, output_padding or dilation is a tuple whose length is not equal to 2.

Supported Platforms:

Ascend

Examples

>>> import mindspore
>>> import numpy as np
>>> from mindspore import Tensor, mint
>>> x = Tensor(np.ones([1, 4, 5, 5]), mindspore.float32)
>>> weight = Tensor(np.ones([4, 8, 3, 3]), mindspore.float32)
>>> output = mint.nn.functional.conv_transpose2d(x, weight)
>>> print(output.shape)
(1, 8, 7, 7)