The deduced formulas of Conv1d and ConvTranspose1d

torch.nn.Conv1d(in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=True)

In the simplest case, the output value of the layer with input size (N,Cin,L) and output (N,Cout,Lout) can be precisely described as:(N,Cout,Lout) can be precisely described as:

out(Ni,Coutj)=bias(Coutj)+∑Cin−1k=0weight(Coutj,k)⋆input(Ni,k)

Parameters:

in_channels (int) – Number of channels in the input image out_channels (int) – Number of channels produced by the convolution kernel_size (int or tuple) – Size of the convolving kernel stride (int or tuple, optional) – Stride of the convolution padding (int or tuple, optional) – Zero-padding added to both sides of the input dilation (int or tuple, optional) – Spacing between kernel elements groups (int, optional) – Number of blocked connections from input channels to output channels bias (bool, optional) – If True, adds a learnable bias to the output

Shape:

• Input: (N,Cin,Lin)
  
• Output: (N,Cout,Lout)    where
• Lout=floor((Lin+2∗padding−dilation∗(kernel_size−1)−1)/stride+1)
  

Variables:	weight (Tensor) – the learnable weights of the module of shape (out_channels, in_channels, kernel_size) bias (Tensor) – the learnable bias of the module of shape (out_channels)

Examples:

>>> m = nn.Conv1d(16, 33, 3, stride=2)
>>> input = autograd.Variable(torch.randn(20, 16, 50))
>>> output = m(input)




________________________________________

torch.nn.ConvTranspose1d(in_channels, out_channels, kernel_size, stride=1, padding=0, output_padding=0, groups=1, bias=True, dilation=1)

Parameters:

in_channels (int) – Number of channels in the input image out_channels (int) – Number of channels produced by the convolution kernel_size (int or tuple) – Size of the convolving kernel stride (int or tuple, optional) – Stride of the convolution padding (int or tuple, optional) – Zero-padding added to both sides of the input output_padding (int or tuple, optional) – Zero-padding added to one side of the output groups (int, optional) – Number of blocked connections from input channels to output channels bias (bool, optional) – If True, adds a learnable bias to the output

Shape:

• Input: (N,Cin,Lin)(N,Cin,Lin)
• Output: (N,Cout,Lout) where
• Lout=(Lin−1)∗stride−2∗padding+kernel_size+output_padding
  

Variables:	weight (Tensor) – the learnable weights of the module of shape (in_channels, out_channels, kernel_size[0], kernel_size[1]) bias (Tensor) – the learnable bias of the module of shape (out_channels)

from: http://pytorch.org/docs/0.1.12/nn.html