nn_hardswish: Hardswish module

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

Hardswish module

Description

Applies the hardswish function, element-wise, as described in the paper: Searching for MobileNetV3

Usage

nn_hardswish()

Details

\mbox{Hardswish}(x) = \left\{ \begin{array}{ll} 0 & \mbox{if } x \le -3, \\ x & \mbox{if } x \ge +3, \\ x \cdot (x + 3)/6 & \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()) {
## Not run: 
m <- nn_hardswish()
input <- torch_randn(2)
output <- m(input)

## End(Not run)

}

torch documentation built on June 7, 2023, 6:19 p.m.