neural_network 0.1.3

Modular neural network with an implementation of back propagation learning algorithm.
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

use neuron::{Neuron as N, rand_weight};
use super::Connection;
use connection::Connection as C;
use fast_io::prelude::*;

#[derive(Clone, Debug)]
pub struct Neuron {
    pub bias: f64,
    pub prev_delta: f64,
    pub output: f64,
    pub sum: f64,
    pub connections: Vec<Connection>
}

impl CustomIO for Neuron {
    fn save<T: CopyIO>(&self, f: &mut T) -> Result<()> {
        f.write_copy(&self.bias)?;
        self.connections.save(f)
    }
    fn load<T: CopyIO>(f: &mut T) -> Result<Self> {
        Ok(Neuron {
            bias: f.read_copy()?,
            prev_delta: 0.0,
            output: 0.0,
            sum: 0.0,
            connections: Vec::load(f)?
        })
    }
}

impl PartialEq for Neuron {
    fn eq(&self, rhs: &Self) -> bool {
        self.bias == rhs.bias && self.connections == rhs.connections
    }
}

impl N for Neuron {
    type C = Connection;

    fn new(prev_count: usize) -> Neuron {
        let mut connections: Vec<Connection> = Vec::with_capacity(prev_count);
        for _ in 0..prev_count {
            connections.push(Connection::new(rand_weight()));
        }
        Neuron {
            bias: rand_weight(),
            prev_delta: 0.0,
            output: 0.0,
            sum: 0.0,
            connections
        }
    }

    fn bias(&self) -> f64 { self.bias }
    fn output(&self) -> f64 { self.output }
    fn set_output(&mut self, output: f64) { self.output = output; }
    fn connections(&mut self) -> &mut Vec<Self::C> { &mut self.connections }
}