Skip to contents

Applies a 2D transposed convolution operator over an input image composed of several input planes.


  stride = 1,
  padding = 0,
  output_padding = 0,
  groups = 1,
  bias = TRUE,
  dilation = 1,
  padding_mode = "zeros"



(int): Number of channels in the input image


(int): Number of channels produced by the convolution


(int or tuple): Size of the convolving kernel


(int or tuple, optional): Stride of the convolution. Default: 1


(int or tuple, optional): dilation * (kernel_size - 1) - padding zero-padding will be added to both sides of each dimension in the input. Default: 0


(int or tuple, optional): Additional size added to one side of each dimension in the output shape. Default: 0


(int, optional): Number of blocked connections from input channels to output channels. Default: 1


(bool, optional): If True, adds a learnable bias to the output. Default: True


(int or tuple, optional): Spacing between kernel elements. Default: 1


(string, optional): 'zeros', 'reflect', 'replicate' or 'circular'. Default: 'zeros'


This module can be seen as the gradient of Conv2d with respect to its input. It is also known as a fractionally-strided convolution or a deconvolution (although it is not an actual deconvolution operation).

  • stride controls the stride for the cross-correlation.

  • padding controls the amount of implicit zero-paddings on both sides for dilation * (kernel_size - 1) - padding number of points. See note below for details.

  • output_padding controls the additional size added to one side of the output shape. See note below for details.

  • dilation controls the spacing between the kernel points; also known as the à trous algorithm. It is harder to describe, but this link_ has a nice visualization of what dilation does.

  • groups controls the connections between inputs and outputs. in_channels and out_channels must both be divisible by groups. For example,

    • At groups=1, all inputs are convolved to all outputs.

    • At groups=2, the operation becomes equivalent to having two conv layers side by side, each seeing half the input channels, and producing half the output channels, and both subsequently concatenated.

    • At groups= in_channels, each input channel is convolved with its own set of filters (of size \(\left\lfloor\frac{out\_channels}{in\_channels}\right\rfloor\)).

The parameters kernel_size, stride, padding, output_padding can either be:

  • a single int -- in which case the same value is used for the height and width dimensions

  • a tuple of two ints -- in which case, the first int is used for the height dimension, and the second int for the width dimension


Depending of the size of your kernel, several (of the last) columns of the input might be lost, because it is a valid cross-correlation_, and not a full cross-correlation. It is up to the user to add proper padding.

The padding argument effectively adds dilation * (kernel_size - 1) - padding amount of zero padding to both sizes of the input. This is set so that when a nn_conv2d and a nn_conv_transpose2d are initialized with same parameters, they are inverses of each other in regard to the input and output shapes. However, when stride > 1, nn_conv2d maps multiple input shapes to the same output shape. output_padding is provided to resolve this ambiguity by effectively increasing the calculated output shape on one side. Note that output_padding is only used to find output shape, but does not actually add zero-padding to output.

In some circumstances when using the CUDA backend with CuDNN, this operator may select a nondeterministic algorithm to increase performance. If this is undesirable, you can try to make the operation deterministic (potentially at a performance cost) by setting torch.backends.cudnn.deterministic = TRUE.


  • Input: \((N, C_{in}, H_{in}, W_{in})\)

  • Output: \((N, C_{out}, H_{out}, W_{out})\) where $$ H_{out} = (H_{in} - 1) \times \mbox{stride}[0] - 2 \times \mbox{padding}[0] + \mbox{dilation}[0] \times (\mbox{kernel\_size}[0] - 1) + \mbox{output\_padding}[0] + 1 $$ $$ W_{out} = (W_{in} - 1) \times \mbox{stride}[1] - 2 \times \mbox{padding}[1] + \mbox{dilation}[1] \times (\mbox{kernel\_size}[1] - 1) + \mbox{output\_padding}[1] + 1 $$


  • weight (Tensor): the learnable weights of the module of shape \((\mbox{in\_channels}, \frac{\mbox{out\_channels}}{\mbox{groups}},\) \(\mbox{kernel\_size[0]}, \mbox{kernel\_size[1]})\). The values of these weights are sampled from \(\mathcal{U}(-\sqrt{k}, \sqrt{k})\) where \(k = \frac{groups}{C_{\mbox{out}} * \prod_{i=0}^{1}\mbox{kernel\_size}[i]}\)

  • bias (Tensor): the learnable bias of the module of shape (out_channels) If bias is True, then the values of these weights are sampled from \(\mathcal{U}(-\sqrt{k}, \sqrt{k})\) where \(k = \frac{groups}{C_{\mbox{out}} * \prod_{i=0}^{1}\mbox{kernel\_size}[i]}\)


if (torch_is_installed()) {
# With square kernels and equal stride
m <- nn_conv_transpose2d(16, 33, 3, stride = 2)
# non-square kernels and unequal stride and with padding
m <- nn_conv_transpose2d(16, 33, c(3, 5), stride = c(2, 1), padding = c(4, 2))
input <- torch_randn(20, 16, 50, 100)
output <- m(input)
# exact output size can be also specified as an argument
input <- torch_randn(1, 16, 12, 12)
downsample <- nn_conv2d(16, 16, 3, stride = 2, padding = 1)
upsample <- nn_conv_transpose2d(16, 16, 3, stride = 2, padding = 1)
h <- downsample(input)
output <- upsample(h, output_size = input$size())
#> [1]  1 16 12 12