1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47

use ndarray::prelude::*;

/// Applies the sigmoid logistic function
#[inline]
pub fn sigmoid(x: f32) -> f32 {
    1.0 / (1.0 + (-x).exp())
}

/// Derivative of the sigmoid function
#[inline]
pub fn sigmoid_prime(x: f32) -> f32 {
    sigmoid(x) * (1.0 - sigmoid(x))
}

/// REctified Linear Unit function
#[inline]
pub fn relu(x: f32) -> f32 {
    if x < 0.0 {
        0.0
    } else {
        x
    }
}

/// Derivative of the ReLU function
#[inline]
pub fn relu_prime(x: f32) -> f32 {
    if x < 0.0 {
        0.0
    } else {
        1.
    }
}

/// Applies the softmax function in-place on a mutable two-dimensional array, making sure that every row has a proportional value that sums to 1.0
#[inline]
pub fn softmax(array: &mut Array2<f32>) {
    for j in 0..array.nrows() {
        let mut sum = 0.;
        for i in 0..array.ncols() {
            sum += array[[j, i]];
        }
        for i in 0..array.ncols() {
            array[[j, i]] = array[[j, i]] / sum;
        }
    }
}