Table Of Contents
Table Of Contents

mx.nd.hard.sigmoid

Description

Computes hard sigmoid of x element-wise.

\[y = max(0, min(1, alpha * x + beta))\]

Arguments

Argument

Description

data

NDArray-or-Symbol.

The input array.

alpha

float, optional, default=0.2.

Slope of hard sigmoid

beta

float, optional, default=0.5.

Bias of hard sigmoid.