snail_nn 0.1.0

small neural network libary, running on the cpu with parallelized stochastic gradient descent
Documentation 
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
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79

use crate::mat::MatF64;

pub struct TrainingBatch {
    pub input: MatF64,
    pub expected: MatF64,
    index : usize,
}

impl TrainingBatch {

    pub fn empty(inpust_size: usize, output_size: usize) -> TrainingBatch {
        TrainingBatch {
            input: MatF64::empty(0, inpust_size),
            expected: MatF64::empty(0, output_size),
            index: 0,
        }
    }

    pub fn add(&mut self, input: &[f64], expected: &[f64]) {

        assert_eq!(input.len(), self.input.cols());
        assert_eq!(expected.len(), self.expected.cols());

        self.input.add_row(input);
        self.expected.add_row(expected);
    }

    pub fn new(input: MatF64, expected: MatF64) -> TrainingBatch {
        assert_eq!(input.rows(), expected.rows());
        TrainingBatch { input, expected, index: 0 }
    }

    pub fn iter(&self) -> impl Iterator<Item = (&[f64], &[f64])> {
        self.input.iter_rows().zip(self.expected.iter_rows())
    }

    pub fn len(&self) -> usize {
        self.input.len()
    }

    pub fn next_chunk(&mut self, size: usize) -> Self {

        let mut input = MatF64::empty(0, self.input.cols());
        let mut expected = MatF64::empty(0, self.expected.cols());

        for i in 0..size {

            let offset = rand::random::<usize>() % 3;
            let index = ( self.index + i + offset) % self.input.rows();

            input.add_row(self.input.get_row(index));
            expected.add_row(self.expected.get_row(index));

            self.index += 1;
        }

        TrainingBatch::new(input, expected)
    }

    pub fn random_chunk(&self, size: usize) -> Self {

        let mut input = MatF64::empty(0, self.input.cols());
        let mut expected = MatF64::empty(0, self.expected.cols());

        let offset = rand::random::<usize>();


        for i in 0..size {

            let index = ( offset + i ) % self.input.rows();

            input.add_row(self.input.get_row(index));
            expected.add_row(self.expected.get_row(index));

        }

        TrainingBatch::new(input, expected)
    }
}