neural-network/Layers.java at main · SparshSuman2005/neural-network · GitHub

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
public class Layers {
    double[][] A;
    double[][] A_prev;
    double[][] w;
    double[][] z;
    double[][] b;
    int n ;
    double[][] dz;
    double[][] db;
    double[][] dw;


    Layers( int n , int n_prev){

        this.n = n;

        this.w = Initializers.xavieruniforminitialize( n , n_prev );
        this.b = Initializers.biasinitialize(n,1);


    }


    public double[][] forward(double[][] A_prev){

        this.A_prev = A_prev;
        double[][] a = Matrixutilities.matmul(w,A_prev);
        this.z = Matrixutilities.add(a, b);
        this.A = Matrixutilities.map(z);

        return this.A;
        }


    public void backpropagate(double[][] dz_next , double[][] w_next ){


        this.dz = Matrixutilities.hadamard((Matrixutilities.matmul(Matrixutilities.transpose(w_next), dz_next)),  Matrixutilities.mapsigmoidderivative(z) );
        this.dw = Matrixutilities.matmul(dz, Matrixutilities.transpose(this.A_prev));
        this.db = dz;


    }

    public void backpropagate_initial(double[][] Y ){

        this.dz = Matrixutilities.substract(A, Y)  ;
        this.dw=Matrixutilities.matmul(dz, Matrixutilities.transpose(this.A_prev));
        this.db = dz;

    }


    public void update(){

        double lr = 0.01;


        this.w = Matrixutilities.substract(this.w, Matrixutilities.consmul(this.dw, lr));
        this.b = Matrixutilities.substract(this.b, Matrixutilities.consmul(this.db, lr));


    }

}