MaxUnpool1d#

classtorch.nn.modules.pooling.MaxUnpool1d(kernel_size,stride=None,padding=0)[source]#

Computes a partial inverse ofMaxPool1d.

MaxPool1d is not fully invertible, since the non-maximal values are lost.

MaxUnpool1d takes in as input the output ofMaxPool1dincluding the indices of the maximal values and computes a partial inversein which all non-maximal values are set to zero.

Note

This operation may behave nondeterministically when the input indices has repeat values.Seepytorch/pytorch#80827 andReproducibility for more information.

Note

MaxPool1d can map several input sizes to the same outputsizes. Hence, the inversion process can get ambiguous.To accommodate this, you can provide the needed output sizeas an additional argumentoutput_size in the forward call.See the Inputs and Example below.

Parameters

• kernel_size (int ortuple) – Size of the max pooling window.

• stride (int ortuple) – Stride of the max pooling window.It is set tokernel_size by default.

• padding (int ortuple) – Padding that was added to the input

Inputs:

• input: the input Tensor to invert

• indices: the indices given out byMaxPool1d

• output_size (optional): the targeted output size

Shape:

• Input:$(N,C,H_{in})$(N,C,Hin​) or$(C,H_{in})$(C,Hin​).

• Output:$(N,C,H_{out})$(N,C,Hout​) or$(C,H_{out})$(C,Hout​), where

  $$H_{out}=(H_{in}-1)\times \text{stride}[0]-2\times \text{padding}[0]+\text{kernel\_size}[0]$$Hout​=(Hin​−1)×stride[0]−2×padding[0]+kernel_size[0]

  or as given byoutput_size in the call operator

Example:

>>>pool=nn.MaxPool1d(2,stride=2,return_indices=True)>>>unpool=nn.MaxUnpool1d(2,stride=2)>>>input=torch.tensor([[[1.,2,3,4,5,6,7,8]]])>>>output,indices=pool(input)>>>unpool(output,indices)tensor([[[ 0.,  2.,  0.,  4.,  0.,  6.,  0., 8.]]])>>># Example showcasing the use of output_size>>>input=torch.tensor([[[1.,2,3,4,5,6,7,8,9]]])>>>output,indices=pool(input)>>>unpool(output,indices,output_size=input.size())tensor([[[ 0.,  2.,  0.,  4.,  0.,  6.,  0., 8.,  0.]]])>>>unpool(output,indices)tensor([[[ 0.,  2.,  0.,  4.,  0.,  6.,  0., 8.]]])

forward(input,indices,output_size=None)[source]#

Runs the forward pass.

Return type

Tensor