test/test_models_torch.py at main · qryptum/test · 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
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
import torch
import torch.nn as nn
import torch.nn.functional as F


class LinearRegression(nn.Module):
    def __init__(self, linear1_weights, linear1_bias):
        super(LinearRegression, self).__init__()

        self.linear1_weights = nn.Parameter(linear1_weights)
        self.linear1_bias = nn.Parameter(linear1_bias)

    def forward(self, x):
        output = F.linear(x, self.linear1_weights.T, self.linear1_bias)
        return output


class MSELoss(nn.Module):
    def __init__(self):
        super(MSELoss, self).__init__()

    def forward(self, output, target):
        loss = F.mse_loss(output, target)
        return loss


class CrossEntropyLoss(nn.Module):
    def __init__(self):
        super(CrossEntropyLoss, self).__init__()

    def forward(self, output, target):
        loss = -torch.sum(target * torch.log(output)) / output.size(0)
        return loss


# Implement the class for crossentropy loss with logsoftmax
class CrossEntropyLoss2(nn.Module):
    def __init__(self):
        super(CrossEntropyLoss2, self).__init__()

    def forward(self, output, target):
        loss = -torch.sum(target * F.log_softmax(output, dim=1)) / output.size(0)
        return loss


class CNN(nn.Module):
    def __init__(
        self,
        conv1_weights,
        conv1_bias,
        conv2_weights,
        conv2_bias,
        linear1_weights,
        linear1_bias,
    ):
        super(CNN, self).__init__()

        self.conv1_weights = nn.Parameter(conv1_weights)
        self.conv1_bias = nn.Parameter(conv1_bias)
        self.conv2_weights = nn.Parameter(conv2_weights)
        self.conv2_bias = nn.Parameter(conv2_bias)
        self.linear1_weights = nn.Parameter(linear1_weights)
        self.linear1_bias = nn.Parameter(linear1_bias)

    def forward(self, x):
        x = F.conv2d(x, self.conv1_weights, self.conv1_bias, stride=1, padding=2)
        x = F.relu(x)
        x = F.max_pool2d(x, 2)
        x = F.conv2d(x, self.conv2_weights, self.conv2_bias, stride=1, padding=2)
        x = F.relu(x)
        x = F.max_pool2d(x, 2)
        x = x.view(x.size(0), -1)
        output = F.linear(x, self.linear1_weights.T, self.linear1_bias)
        return output

    def print_grads(self):
        print("\nCONV1 WEIGHTS", self.conv1_weights.grad.shape)
        print(self.conv1_weights.grad)

        print("\nCONV1 BIAS", self.conv1_bias.grad.shape)
        print(self.conv1_bias.grad)

        print("\nCONV2 WEIGHTS", self.conv2_weights.grad.shape)
        print(self.conv2_weights.grad)

        print("\nCONV2 BIAS", self.conv2_bias.grad.shape)
        print(self.conv2_bias.grad)

        print("\nLINEAR1 WEIGHTS", self.linear1_weights.grad.shape)
        print(self.linear1_weights.grad)

        print("\nLINEAR1 BIAS", self.linear1_bias.grad.shape)
        print(self.linear1_bias.grad)


class SimpleNN(nn.Module):
    def __init__(
        self,
        linear1_weights,
        linear1_bias,
        linear2_weights,
        linear2_bias,
        linear3_weights,
        linear3_bias,
    ):
        super(SimpleNN, self).__init__()

        self.linear1_weights = nn.Parameter(linear1_weights)
        self.linear1_bias = nn.Parameter(linear1_bias)
        self.linear2_weights = nn.Parameter(linear2_weights)
        self.linear2_bias = nn.Parameter(linear2_bias)
        self.linear3_weights = nn.Parameter(linear3_weights)
        self.linear3_bias = nn.Parameter(linear3_bias)

        self.relu1 = nn.ReLU()
        self.relu2 = nn.ReLU()

    def forward(self, x):
        x1 = F.linear(x, self.linear1_weights.T, self.linear1_bias)
        x2 = self.relu1(x1)
        x3 = F.linear(x2, self.linear2_weights.T, self.linear2_bias)
        x4 = self.relu2(x3)
        y_pred = F.linear(x4, self.linear3_weights.T, self.linear3_bias)
        return y_pred