Applies the element-wise function: $$ \mbox{LogSigmoid}(x) = \log\left(\frac{ 1 }{ 1 + \exp(-x)}\right) $$
Shape
Input: \((N, *)\) where
*
means, any number of additional dimensionsOutput: \((N, *)\), same shape as the input
Examples
if (torch_is_installed()) {
m <- nn_log_sigmoid()
input <- torch_randn(2)
output <- m(input)
}