jiro_nn 0.8.1

Neural Networks framework with model building & data preprocessing features.
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
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138

use std::cmp::Ordering;

use rand::Rng;

use crate::layer::{DropoutLayer, LearnableLayer, ParameterableLayer};
use crate::linalg::Scalar;
use crate::vision::conv_network::ConvNetworkLayer;

use super::{ConvLayer, Image};
use crate::vision::image::ImageTrait;
use crate::vision::conv_activation::ConvActivationLayer;
use crate::vision::image_layer::ImageLayer;

#[derive(Debug)]
pub struct FullConvLayer {
    conv: Box<dyn ConvLayer>,
    activation: ConvActivationLayer,
    dropout_enabled: bool,
    dropout_rate: Option<Scalar>,
    mask: Option<Image>,
}

impl FullConvLayer {
    pub fn new(
        conv: Box<dyn ConvLayer>,
        activation: ConvActivationLayer,
        dropout: Option<Scalar>,
    ) -> Self {
        Self {
            conv,
            activation,
            dropout_rate: dropout,
            dropout_enabled: false,
            mask: None,
        }
    }

    fn generate_dropout_mask(
        &mut self,
        kern_size: (usize, usize, usize),
        nkern: usize,
    ) -> Option<(Image, Scalar)> {
        if let Some(dropout_rate) = self.dropout_rate {
            let mut rng = rand::thread_rng();
            let dropout_mask = Image::from_fn(
                kern_size.0,
                kern_size.1,
                kern_size.2,
                nkern,
                |_, _, _, _| {
                    if rng
                        .gen_range((0.0 as Scalar)..(1.0 as Scalar))
                        .total_cmp(&self.dropout_rate.unwrap())
                        == Ordering::Greater
                    {
                        1.0
                    } else {
                        0.0
                    }
                },
            );
            Some((dropout_mask, dropout_rate))
        } else {
            None
        }
    }
}

impl ImageLayer for FullConvLayer {
    fn forward(&mut self, mut input: Image) -> Image {
        let output = if self.dropout_enabled {
            if let Some((mask, _)) = self.generate_dropout_mask(input.image_dims(), input.samples())
            {
                input = input.component_mul(&mask);
                self.mask = Some(mask);
            };
            self.conv.forward(input)
        } else {
            if let Some(dropout_rate) = self.dropout_rate {
                self.conv.scale_kernels(1.0 - dropout_rate);
                let output = self.conv.forward(input);
                self.conv.scale_kernels(1.0 / (1.0 - dropout_rate));
                output
            } else {
                self.conv.forward(input)
            }
        };

        self.activation.forward(output)
    }

    fn backward(&mut self, epoch: usize, output_gradient: Image) -> Image {
        let activation_input_gradient = self.activation.backward(epoch, output_gradient);
        let input_gradient = self.conv.backward(epoch, activation_input_gradient);

        if let Some(mask) = &self.mask {
            input_gradient.component_mul(&mask)
        } else {
            input_gradient
        }
    }
}

impl LearnableLayer for FullConvLayer {
    fn get_learnable_parameters(&self) -> Vec<Vec<Scalar>> {
        self.conv.get_learnable_parameters()
    }

    fn set_learnable_parameters(&mut self, params_matrix: &Vec<Vec<Scalar>>) {
        self.conv.set_learnable_parameters(params_matrix)
    }
}

impl DropoutLayer for FullConvLayer {
    fn enable_dropout(&mut self) {
        self.dropout_enabled = true;
    }

    fn disable_dropout(&mut self) {
        self.dropout_enabled = false;
    }
}

impl ParameterableLayer for FullConvLayer {
    fn as_learnable_layer(&self) -> Option<&dyn crate::layer::LearnableLayer> {
        Some(self)
    }

    fn as_learnable_layer_mut(&mut self) -> Option<&mut dyn crate::layer::LearnableLayer> {
        Some(self)
    }

    fn as_dropout_layer(&mut self) -> Option<&mut dyn crate::layer::DropoutLayer> {
        Some(self)
    }
}

impl ConvNetworkLayer for FullConvLayer {}