MSAT/ImageNet_ResNet2.py at main · Brain-Cog-Lab/MSAT · 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
125
126
127
128
import torch
import torch.nn as nn
import torch.nn.functional as F
import torchvision.models as models
import torchvision.transforms as transforms
import torchvision.datasets as datasets
import matplotlib
# matplotlib.use('agg')
import matplotlib.pyplot as plt
import numpy as np
from tqdm import tqdm
import time
import os
from utils import load_imagenet, evaluate_accuracy
from pytorchcv.model_provider import get_model as ptcv_get_model


device = torch.device('cuda:1' if torch.cuda.is_available() else 'cpu')
max_act = []
resnet34 = ((3, 64, 64), (4, 64, 128), (6, 128, 256), (3, 256, 512))


class ConvBlock(nn.Module):
    def __init__(self, in_channels, out_channels, norelu=False, stride=1, kernel=3, pad=1):
        super(ConvBlock, self).__init__()
        self.conv = nn.Conv2d(in_channels, out_channels, kernel_size=kernel, stride=stride, padding=pad, bias=False)
        self.bn = nn.BatchNorm2d(out_channels)
        self.norelu = norelu

        if not norelu:
            self.activ = nn.ReLU(inplace=True)

    def forward(self, x):
        if not self.norelu:
            return self.activ(self.bn(self.conv(x)))
        else:
            return self.bn(self.conv(x))


class ResInitBlock(nn.Module):
    def __init__(self):
        super(ResInitBlock, self).__init__()
        self.conv = ConvBlock(3, 64, kernel=7, stride=2, pad=3)
        self.pool = nn.MaxPool2d(3,2,1)

    def forward(self, x):
        return self.pool(self.conv(x))


class ResBlock(nn.Module):
    def __init__(self, inchannel, outchannel, stride=1, pad=1):
        super(ResBlock, self).__init__()
        self.conv1 = ConvBlock(inchannel, outchannel, stride=stride, pad=pad)
        self.conv2 = ConvBlock(outchannel, outchannel, True)

    def forward(self, x):
        return self.conv2(self.conv1(x))


class Scale():
    def __init__(self):
        self.scale = 1.0

    def __call__(self, x):
        return self.scale * x


class ResUnit(nn.Module):
    def __init__(self, inchannel, outchannel, stride=1, pad=1):
        super(ResUnit, self).__init__()
        self.inchannel = inchannel
        self.outchannel = outchannel
        self.body = ResBlock(inchannel, outchannel, stride=stride, pad=1)

        if self.inchannel == self.outchannel:
            self.identity_conv = Scale()
        else:
            self.identity_conv = ConvBlock(inchannel, outchannel, kernel=1, stride=2, pad=pad, norelu=True)
        self.activ = nn.ReLU(True)

    def forward(self, x):
        return self.activ(self.body(x)+self.identity_conv(x))


class CNN(nn.Module):
    def __init__(self, arch=resnet34):
        super(CNN, self).__init__()
        self.hooks = []
        features = nn.Sequential()
        features.add_module('init_block', ResInitBlock())

        for i, (num_residual, in_channel, out_channel) in enumerate(arch):
            blk = nn.Sequential()
            for j in range(num_residual):
                if j == 0 and i!=0:
                    blk.add_module('unit%d'%(j+1), ResUnit(in_channel, out_channel, stride=2, pad=0))
                elif j==0 and i == 0:
                    blk.add_module('unit%d' % (j + 1), ResUnit(in_channel, out_channel, stride=1, pad=0))
                else:
                    blk.add_module('unit%d'%(j+1), ResUnit(out_channel, out_channel))
            features.add_module('stage%d'%(i+1), blk)

        self.features = features
        self.finalpool = nn.AvgPool2d(kernel_size=7, stride=1)
        self.output = nn.Linear(512, 1000)

    def forward(self, X):
        X = self.features(X)
        X = self.finalpool(X)
        out = self.output(X.view(X.shape[0], -1))
        return out


if __name__ == '__main__':
    train_iter, test_iter, _, _ = load_imagenet(root='./data/ImageNet',batch_size=64)

    # net = CNN()
    net = ptcv_get_model('resnet34', pretrained=False)
    print(net)
    # [net.hooks[i].remove() for i in range(len(net.hooks))]
    dict = ptcv_get_model('resnet34', pretrained=True).to(device).state_dict()
    net.load_state_dict(dict)
    net.eval()

    # net = nn.DataParallel(net, device_ids = [0, 1, 2, 3])
    net = net.to(device)
    acc = evaluate_accuracy(test_iter, net, device)
    print(acc)