[Pytorch] VGG 구현 및 정리

VGG 구현 및 정리

모두의 딥러닝 시즌2 - Pytorch를 참고 했습니다.
모두의 딥러닝 시즌2 깃헙

---

import torch.nn as nn

class VGG(nn.Module):
    def __init__(self, features, num_classes=1000, init_weights=True):
        super(VGG, self).__init__()

        self.features = features
        self.avgpool = nn.AdaptiveAvgPool2d(7)
        self.classifier = nn.Sequential(
            nn.Linear(512*7*7, 4096),
            nn.ReLU(True),
            nn.Dropout(),
            nn.Linear(4096, 4096),
            nn.ReLU(True),
            nn.Dropout(),
            nn.Linear(4096, num_classes),
        )

        if init_weights:
            self._initialize_weights()

    def forward(self, x):
        x = self.features(x)
        x = self.avgpool(x)
        x = x.view(s.size(0), -1)
        x = self.classifier(x)
        return x

    def _initialize_weights(self):
        for m in self.modules():
            if isinstance(m, nn.Conv2d):
                nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
                if m.bias is not None:
                    nn.init.constant_(m.bias, 0)
            elif isinstance(m, nn.BatchNorm2d):
                nn.init.constant_(m.weight, 1)
                nn.init.constant_(m.bias, 0)
            elif isinstance(m, nn.Linear):
                nn.init.normal_(m.weight, 0, 0.01)
                nn.init.constant_(m.bias, 0)

• pytorch에서 class 형태의 모델은 항상 nn.Module을 상속 받아야 하며, super(모델명, self).init()을 통해 nn.Module.init()을 실행시키는 코드가 필요하다
• forward()는 모델이 학습 데이터를 입력 받아서 forward prop을 진행시키는 함수
• VGG의 여러 모델간(VGG16, VGG19...)의 호환성을 위해, 가변적인 부분인 features은 입력으로 받고, 나머지 고정된 부분을 class 내에 설계한다.
• self.modules() -> 모델 클래스에서 정의된 layer들을 iterable로 차례로 반환
• isinstance() -> 차례로 layer을 입력하여, layer의 형태를 반환(nn.Conv2d, nn.BatchNorm2d ...)
• nn.init.kaiming_normal -> he initialization의 한 종류
  

  

• torch.nn.init.constant_(tensor, val) -> tensor을 val로 초기화
• torch.nn.init.normal_(tensor, mean=0.0, std=1.0) -> tensor을 mean, std의 normal distrubution으로 초기화

def make_layers(cfg, batch_norm=False):
    layers = []
    in_channels = 3

    for v in cfg:
        if v == 'M':# max pooling
            layers += [nn.MaxPool2d(kernel_size=2, stride=2)]
        else:
            conv2d = nn.Conv2d(in_channels, v, kernel_size=3, padding=1)
            if batch_norm:
                layers += [conv2d, nn.BatchNorm2d(v), nn.ReLU(inplace=True)]
            else:
                layers += [conv2d, nn.ReLU(inplace=True)]
            in_channels = v

    return nn.Sequential(*layers)

• cfg의 입력값에 따라 layers에 차례로 쌓아감
• 숫자(output channel)가 들어올 경우, conv2d 수행, output channel이 input channel이 됨
• M이 들어올 경우, max pooling 수행

cfg = {
    #8 + 3 =11 == vgg11
    'A': [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'], 
    # 10 + 3 = vgg 13
    'B': [64, 64, 'M', 128, 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'], 
    #13 + 3 = vgg 16
    'D': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512, 'M'], 
    # 16 +3 =vgg 19
    'E': [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 256, 'M', 512, 512, 512, 512, 'M', 512, 512, 512, 512, 'M']
}

https://medium.com/coinmonks/paper-review-of-vggnet-1st-runner-up-of-ilsvlc-2014-image-classification-d02355543a11

feature = make_layers(cfg['A'], batch_norm=True)

CNN = VGG(feature, num_classes=10, init_weights=True)

CNN

VGG(
(features): Sequential(
(0): Conv2d(3, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(2): ReLU(inplace=True)
(3): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)

(4): Conv2d(64, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(5): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(6): ReLU(inplace=True)
(7): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)

(8): Conv2d(128, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(9): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(10): ReLU(inplace=True)
(11): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(12): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(13): ReLU(inplace=True)
(14): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)

(15): Conv2d(256, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(16): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(17): ReLU(inplace=True)
(18): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(19): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(20): ReLU(inplace=True)
(21): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)

(22): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(23): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(24): ReLU(inplace=True)
(25): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))
(26): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(27): ReLU(inplace=True)
(28): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
)

(avgpool): AdaptiveAvgPool2d(output_size=7)

(classifier): Sequential(
(0): Linear(in_features=25088, out_features=4096, bias=True)
(1): ReLU(inplace=True)
(2): Dropout(p=0.5, inplace=False)
(3): Linear(in_features=4096, out_features=4096, bias=True)
(4): ReLU(inplace=True)
(5): Dropout(p=0.5, inplace=False)
(6): Linear(in_features=4096, out_features=10, bias=True)
)
)

VGG11의 형태 아래 그림과 비교해보자

현재 VGG의 target이 CIFAR-10이기 때문에, 마지막의 FC layer은 이미지넷의 1000이 아닌, 10으로 설정하였다.