nn_lp_pool2d: Applies a 2D power-average pooling over an input signal...

Description Usage Arguments Details Shape Note Examples

Description

On each window, the function computed is:

f(X) = √[p]{∑_{x \in X} x^{p}}

Usage

1
nn_lp_pool2d(norm_type, kernel_size, stride = NULL, ceil_mode = FALSE)

Arguments

norm_type

if inf than one gets max pooling if 0 you get sum pooling ( proportional to the avg pooling)

kernel_size

the size of the window

stride

the stride of the window. Default value is kernel_size

ceil_mode

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

Details

The parameters kernel_size, stride can either be:

Shape

H_{out} = ≤ft\lfloor\frac{H_{in} - \mbox{kernel\_size}[0]}{\mbox{stride}[0]} + 1\right\rfloor

W_{out} = ≤ft\lfloor\frac{W_{in} - \mbox{kernel\_size}[1]}{\mbox{stride}[1]} + 1\right\rfloor

Note

If the sum to the power of p is zero, the gradient of this function is not defined. This implementation will set the gradient to zero in this case.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
if (torch_is_installed()) {
  
# power-2 pool of square window of size=3, stride=2
m <- nn_lp_pool2d(2, 3, stride=2)
# pool of non-square window of power 1.2
m <- nn_lp_pool2d(1.2, c(3, 2), stride=c(2, 1))
input <- torch_randn(20, 16, 50, 32)
output <- m(input)

}

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