softmax: Number Softmax

In modello: Homemade Deep Learning Library

Description

Calculates the Softmax function

sofmax = \frac{exp(\bf x)}{∑{exp(\bf x)}}

Usage

## S3 method for class 'modello_number'
softmax(x, k = 0)

## S3 method for class 'numeric'
softmax(x, k = 0)

## S3 method for class 'matrix'
softmax(x, k = 0)

softmax(x, k)

Arguments

x	a 'numeric' or a reference object of class 'number'
k	dimension index. Currently constained to be <= 2

Value

Returns a 'numeric' or reference object of class 'number'

Author(s)

Filippo Monari

Examples

modello.init(10, 10, 10, 10)
## For modello_number matrices
x = number(matrix(rnorm(9), 3, 3))
y1 = softmax(x)
print(y1)
print(y1$v)
y2 = softmax(x, 1)
print(y2)
print(y2$v)
y3 = softmax(x, 2)
print(y3)
print(y3$v)
modello.close()
## For numerics
x = rnorm(5)
y = softmax(x)
print(y)
## For matrices
x = matrix(rnorm(9), 3, 3)
y1 = softmax(x)
print(y1)
y2 = softmax(x, 1)
print(y2)
y3 = softmax(x, 2)
print(y3)

modello documentation built on Feb. 2, 2021, 9:06 a.m.