sparse_plus

Contents

sparse_plus#

class brainunit.math.sparse_plus(x, unit_to_scale=None)#

Sparse plus function.

Computes the function:

\[\begin{split}\mathrm{sparse\_plus}(x) = \begin{cases} 0, & x \leq -1\\ \frac{1}{4}(x+1)^2, & -1 < x < 1 \\ x, & 1 \leq x \end{cases}\end{split}\]

This is the twin function of the softplus activation ensuring a zero output for inputs less than -1 and a linear output for inputs greater than 1, while remaining smooth, convex, monotonic by an adequate definition between -1 and 1.

Parameters:

  • x (Quantity | Array | ndarray | bool | number | bool | int | float | complex) – Input array. Must be unitless if a Quantity.

  • unit_to_scale (Unit | None) – Unit used to convert x to a dimensionless number before applying the activation.

Returns:

out – An array with the same shape as x.

Return type:

Array

Examples

>>> import jax.numpy as jnp
>>> import saiunit.math as sumath
>>> sumath.sparse_plus(jnp.array([-2., 0., 2.]))
Array([0.  , 0.25, 2.  ], dtype=float32)
Contents