nn_max_pool1d: MaxPool1D module

Description Usage Arguments Details Shape Examples

Description

Applies a 1D max pooling over an input signal composed of several input planes.

Usage

1
2
3
4
5
6
7
8
nn_max_pool1d(
  kernel_size,
  stride = NULL,
  padding = 0,
  dilation = 1,
  return_indices = FALSE,
  ceil_mode = FALSE
)

Arguments

kernel_size

the size of the window to take a max over

stride

the stride of the window. Default value is kernel_size

padding

implicit zero padding to be added on both sides

dilation

a parameter that controls the stride of elements in the window

return_indices

if TRUE, will return the max indices along with the outputs. Useful for nn_max_unpool1d() later.

ceil_mode

when TRUE, will use ceil instead of floor to compute the output shape

Details

In the simplest case, the output value of the layer with input size (N, C, L) and output (N, C, L_{out}) can be precisely described as:

out(N_i, C_j, k) = \max_{m=0, …, \mbox{kernel\_size} - 1} input(N_i, C_j, stride \times k + m)

If padding is non-zero, then the input is implicitly zero-padded on both sides for padding number of points. dilation controls the spacing between the kernel points. It is harder to describe, but this link has a nice visualization of what dilation does.

Shape

L_{out} = ≤ft\lfloor \frac{L_{in} + 2 \times \mbox{padding} - \mbox{dilation} \times (\mbox{kernel\_size} - 1) - 1}{\mbox{stride}} + 1\right\rfloor

Examples

1
2
3
4
5
6
7
if (torch_is_installed()) {
# pool of size=3, stride=2
m <- nn_max_pool1d(3, stride=2)
input <- torch_randn(20, 16, 50)
output <- m(input)

}

torch documentation built on Oct. 7, 2021, 9:22 a.m.