LogSigmoid module — nn_log_sigmoid • torch

Applies the element-wise function: $LogSigmoid (x) = \log (\frac{1}{1 + \exp (- x)})$

Usage

nn_log_sigmoid()

Shape

Input: $(N, *)$ where * means, any number of additional dimensions
Output: $(N, *)$ , same shape as the input

Examples

if (torch_is_installed()) {
m <- nn_log_sigmoid()
input <- torch_randn(2)
output <- m(input)
}