nn_softshrink: Softshrink module

In torch: Tensors and Neural Networks with 'GPU' Acceleration

Softshrink module

Description

Applies the soft shrinkage function elementwise:

Usage

nn_softshrink(lambd = 0.5)

Arguments

lambd

the \lambda (must be no less than zero) value for the Softshrink formulation. Default: 0.5

Details

\mbox{SoftShrinkage}(x) = \left\{ \begin{array}{ll} x - \lambda, & \mbox{ if } x > \lambda \\ x + \lambda, & \mbox{ if } x < -\lambda \\ 0, & \mbox{ otherwise } \end{array} \right.

Shape

• Input: (N, *) where * means, any number of additional dimensions

• Output: (N, *), same shape as the input

Examples

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

torch documentation built on May 29, 2024, 9:54 a.m.