yarnn 0.1.0

Yet Another rust Neural Network framework
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
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183

#[allow(dead_code)]
pub fn valid_conv2d_5x5(y: &mut [f32], x: &[f32], w: &[f32], alpha: f32,
                        x_rows: isize, x_cols: isize, s_row: isize, s_col: isize) {
    
    let y_rows = (x_rows - 5) / s_row + 1;
    let y_cols = (x_cols - 5) / s_col + 1;

    let y = &mut y[0..(y_rows * y_cols) as usize];
    let x = &x[0..(x_rows * x_cols) as usize];
    let w = &w[0..25];
    
    for y_y in 0..y_rows {
        for y_x in 0..y_cols {
            let mut xi = s_row * y_y * x_cols + s_col * y_x;            
            let mut sum = 0.0;

            sum += x[(xi + 0) as usize] * w[0];
            sum += x[(xi + 1) as usize] * w[1];
            sum += x[(xi + 2) as usize] * w[2];
            sum += x[(xi + 3) as usize] * w[3];
            sum += x[(xi + 4) as usize] * w[4];
            xi += x_cols;
            
            sum += x[(xi + 0) as usize] * w[5];
            sum += x[(xi + 1) as usize] * w[6];
            sum += x[(xi + 2) as usize] * w[7];
            sum += x[(xi + 3) as usize] * w[8];
            sum += x[(xi + 4) as usize] * w[9];
            xi += x_cols;
            
            sum += x[(xi + 0) as usize] * w[10];
            sum += x[(xi + 1) as usize] * w[11];
            sum += x[(xi + 2) as usize] * w[12];
            sum += x[(xi + 3) as usize] * w[13];
            sum += x[(xi + 4) as usize] * w[14];
            xi += x_cols;
            
            sum += x[(xi + 0) as usize] * w[15];
            sum += x[(xi + 1) as usize] * w[16];
            sum += x[(xi + 2) as usize] * w[17];
            sum += x[(xi + 3) as usize] * w[18];
            sum += x[(xi + 4) as usize] * w[19];
            xi += x_cols;
            
            sum += x[(xi + 0) as usize] * w[20];
            sum += x[(xi + 1) as usize] * w[21];
            sum += x[(xi + 2) as usize] * w[22];
            sum += x[(xi + 3) as usize] * w[23];
            sum += x[(xi + 4) as usize] * w[24];

            y[(y_y * y_cols + y_x) as usize] += alpha * sum;
        }
    }
}


#[allow(dead_code)]
pub fn full_xcorr2d_5x5(y: &mut [f32], x: &[f32], w: &[f32], alpha: f32,
                        x_rows: isize, x_cols: isize, 
                        s_row: isize, s_col: isize) {
    
    let y_cols = (x_cols - 1) * s_col + 5;
    let y_rows = (x_rows - 1) * s_row + 5;

    let y = &mut y[0..(y_rows * y_cols) as usize];
    let x = &x[0..(x_rows * x_cols) as usize];
    let w = &w[0..25];
    
    for x_y in 0..x_rows {
        for x_x in 0..x_cols {
            let mut yi = s_row * x_y * y_cols + s_col * x_x;
            let z = alpha * x[(x_y * x_cols + x_x) as usize];

            y[(yi + 0) as usize] += z * w[24];
            y[(yi + 1) as usize] += z * w[23];
            y[(yi + 2) as usize] += z * w[22];
            y[(yi + 3) as usize] += z * w[21];
            y[(yi + 4) as usize] += z * w[20];
            yi += y_cols;
            
            y[(yi + 0) as usize] += z * w[19];
            y[(yi + 1) as usize] += z * w[18];
            y[(yi + 2) as usize] += z * w[17];
            y[(yi + 3) as usize] += z * w[16];
            y[(yi + 4) as usize] += z * w[15];
            yi += y_cols;
            
            y[(yi + 0) as usize] += z * w[14];
            y[(yi + 1) as usize] += z * w[13];
            y[(yi + 2) as usize] += z * w[12];
            y[(yi + 3) as usize] += z * w[11];
            y[(yi + 4) as usize] += z * w[10];
            yi += y_cols;
            
            y[(yi + 0) as usize] += z * w[9];
            y[(yi + 1) as usize] += z * w[8];
            y[(yi + 2) as usize] += z * w[7];
            y[(yi + 3) as usize] += z * w[6];
            y[(yi + 4) as usize] += z * w[5];
            yi += y_cols;
            
            y[(yi + 0) as usize] += z * w[4];
            y[(yi + 1) as usize] += z * w[3];
            y[(yi + 2) as usize] += z * w[2];
            y[(yi + 3) as usize] += z * w[1];
            y[(yi + 4) as usize] += z * w[0];
        }
    }
}

pub fn conv2d_forward_5x5(y: &mut [f32], x: &[f32], w: &[f32],
                          bs: isize, x_channels: isize, y_channels: isize,
                          x_rows: isize, x_cols: isize, s_row: isize, s_col: isize) {
    
    let y_rows = (x_rows - 5) / s_row + 1;
    let y_cols = (x_cols - 5) / s_col + 1;
    
    let x_img_size = x_rows * x_cols;
    let y_img_size = y_rows * y_cols;
    let w_img_size = 25;
    
    let x_batch_size = x_channels * x_img_size;
    let y_batch_size = y_channels * y_img_size;
    
    let y = &mut y[0..(bs * y_batch_size) as usize];
    let x = &x[0..(bs * x_batch_size) as usize];
    let w = &w[0..(y_channels * w_img_size) as usize];
    
    for bi in 0..bs {
        for x_ch in 0..x_channels {
            let x_offset = (bi * x_batch_size + x_ch * x_img_size) as usize;
            let x_img = &x[x_offset..x_offset + x_img_size as usize];

            for y_ch in 0..y_channels {
                let y_offset = (bi * y_batch_size + y_ch * y_img_size) as usize;
                let y_img = &mut y[y_offset..y_offset + y_img_size as usize];
                
                let w_offset = (y_ch * w_img_size) as usize;
                let w = &w[w_offset..w_offset + w_img_size as usize];
                
                valid_conv2d_5x5(y_img, x_img, w, 1.0, x_rows, x_cols, s_row, s_col);
            }
        }   
    }
} 


pub fn conv2d_backward_5x5(dx: &mut [f32], dy: &[f32], w: &[f32],
                       bs: isize, x_channels: isize, y_channels: isize,
                       y_rows: isize, y_cols: isize,
                       s_row: isize, s_col: isize) {

    let x_cols = (y_cols - 1) * s_col + 5;
    let x_rows = (y_rows - 1) * s_row + 5;
    
    let dx_img_size = x_rows * x_cols;
    let dy_img_size = y_rows * y_cols;
    let w_img_size = 25;
    
    let dx_batch_size = x_channels * dx_img_size;
    let dy_batch_size = y_channels * dy_img_size;
    
    let dx = &mut dx[0..(bs * dx_batch_size) as usize];
    let dy = &dy[0..(bs * dy_batch_size) as usize];
    let w = &w[0..(y_channels * w_img_size) as usize];
    
    for bi in 0..bs {
        for y_ch in 0..y_channels {
            let dy_offset = (bi * dy_batch_size + y_ch * dy_img_size) as usize;
            let dy_img = &dy[dy_offset..dy_offset + dy_img_size as usize];
            
            for x_ch in 0..x_channels {
                let dx_offset = (bi * dx_batch_size + x_ch * dx_img_size) as usize;
                let dx_img = &mut dx[dx_offset..dx_offset + dx_img_size as usize];
                
                let w_offset = (y_ch * w_img_size) as usize;
                let w = &w[w_offset..w_offset + w_img_size as usize];
                
                full_xcorr2d_5x5(dx_img, dy_img, w, 1.0, y_rows, y_cols, s_row, s_col);
            }
        }   
    }
}