Pytorch搭建GoogleNet神经网络-编程知识

一、创建卷积模板文件

因为每次使用卷积层都需要调用Con2d和relu激活函数，每次都调用非常麻烦，就将他们打包在一起写成一个类。

in_channels：输入矩阵深度作为参数输入

out_channels: 输出矩阵深度作为参数输入

经过卷积层和relu激活函数之后通过正向传播得到输出。

class BasicConv2d(nn.Module):def __init__(self, in_chaannels, out_channels, **kwargs):super(BasicConv2d, self).__init__()self.conv = nn.Conv2d(in_chaannels, out_channels, **kwargs)self.conv = nn.ReLU(inplace=True)# 定义正向传播  def forward(self, x):x = self.conv(x)x = self.relu(x)return x

二、定义Inception结构

def __init__(self, in_channels, ch1x1, ch3x3red, ch3x3, ch5x5red, ch5x5, pool_proj):

in_channels：输入特征矩阵的参数

ch1x1, ch3x3red, ch3x3, ch5x5red, ch5x5, pool_proj：表格中对应需要的参数。

branch1、2、3、4 分别对应inception结构中的四个分支，第一层只有一个分支，第二、三、四层都有1x1卷积核起到降维的作用。

所有定义的参数都来自于GoogleNet神经网络的参数表格。

详见GoogleNet神经网络介绍-CSDN博客

class Inception(nn.Module):def __init__(self, in_channels, ch1x1, ch3x3red, ch3x3, ch5x5red, ch5x5, pool_proj):super(Inception, self).__init__()self.branch1 = BasicConv2d(in_channels, ch1x1, kernel_size=1)self.branch2 = nn.Sequential(BasicConv2d(in_channels, ch3x3red, kernel_size=1),BasicConv2d(ch3x3red, ch3x3, kernel_size=3, padding=1)  # padding=1 保证输出大小等于输入大小)self.branch3 = nn.Sequential(BasicConv2d(in_channels, ch5x5red, kernel_size=1),BasicConv2d(ch5x5red, ch5x5, kernel_size=5, padding=2))self.branch4 = nn.Sequential(nn.MaxPool2d(kernel_size=3, stride=1, padding=1),BasicConv2d(in_channels, pool_proj, kernel_size=1))def forward(self, x):  # 将输出的特征矩阵分别输出到branch1234中branch1 = self.branch1(x)branch2 = self.branch2(x)branch3 = self.branch3(x)branch4 = self.branch4(x)# 将输出放入列表中outputs = [branch1, branch2, branch3, branch4]# 对四个输出特征矩阵在channel维度进行合并。“1”指需要合并的维度return torch.cat(outputs, 1)

三、辅助分类器

def __init__(self, in_channels, num_classes)：表示输入特征矩阵的个数和要分类的类别数。

辅助分类器1对应参数：N x 512 x 14 x 14 ,辅助分类器2对应参数：N x 528 x 14 x 14

在经过平均池化下采样后高度和宽度变成了4 x 4

辅助分类器1对应参数：N x 512 x 4 x 4 ,辅助分类器2对应参数：N x 528 x 4 x 4。

x = F.dropout(x, 0.5, training=self.training)

输入特征矩阵按照50%的比例随机失活。

training=self.training

当实例化一个模型model时，可以通过model.train() 和 model.eval()来控制模型的状态，在model.train()模式下，self.training=True, 在model.eval()模式下，self.training=False。

# 辅助分类器
class InceptionAux(nn.Module):def __init__(self, in_channels, num_classes):super(InceptionAux, self).__init__()self.averagePool = nn.AvgPool2d(kernel_size=5, stride=3)  # 平均池化下采样层self.conv = BasicConv2d(in_channels, 128, kernel_size=1)   # 1 x 1 的卷积层# 全连接层self.fc1 = nn.Linear(2048, 1024)self.fc2 = nn.Linear(1024, num_classes)# 定义正向传播def forward(self, x):# aux1：N x 512 x 14 x 14 aux2：N x 528 x 14 x 14x = self.averagePool(x)# aux1：N x 512 x 4 x 4   aux2：N x 528 x 4 x 4。x = self.conv(x)x = torch.flatten(x, 1)    # "1"表示按channel维度展平x = F.dropout(x, 0.5, training=self.training)x = self.fc2(x)return x

四、定义层结构

def __init__(self, num_classes=1000, aux_logits=True, init_weights=False)：

num_classes:分类类别个数

aux_logits: 是否使用辅助分类器

init_weight: 是否为权重进行初始化

self.aux_logits = aux_logits：将变量传入变为类变量

conv1、conv2、conv3 都是使用之前定义的卷积模板文件来定义的。

inception结构使用定义的 Inception结构。

self.avgpool = nn.AdaptiveAvgPool2d((1, 1))：自适应平均池化下采样

括号里输入所需要的输出特征矩阵的高和宽。

使用自适应平均池化下采样的好处：无论输入特征矩阵的高和宽时一个什么样的大小，都能得到所指定的高和宽，这样就可以不用限定输入图像的尺寸了。

class GoogleNet(nn.Module):def __init__(self, num_classes=1000, aux_logits=True, init_weights=False):super(GoogleNet, self).__init__()self.aux_logits = aux_logitsself.conv1 = BasicConv2d(3, 64, kernel_size=7, stride=2, padding=3)self.maxpool1 = nn.MaxPool2d(3, stride=2, ceil_mode=True)self.conv2 = BasicConv2d(64, 64, kernel_size=1)self.conv3 = BasicConv2d(64, 192, kernel_size=3, padding=1)self.maxpool2 = nn.MaxPool2d(3, stride=2, ceil_mode=True)self.inception3a = Inception(192, 64, 96, 128, 16, 32, 32)self.inception3b = Inception(256, 128, 128, 192, 32, 96, 64)self.maxpool3 = nn.MaxPool2d(3, stride=2, ceil_mode=True)self.inception4a = Inception(480, 192, 96, 208, 16, 48, 64)self.inception4b = Inception(512, 160, 112, 224, 24, 64, 64)self.inception4c = Inception(512, 128, 128, 256, 24, 64, 64)self.inception4d = Inception(512, 112, 144, 288, 32, 64, 64)self.inception4e = Inception(528, 256, 160, 320, 32, 128, 128)self.maxpool4 = nn.MaxPool2d(2, stride=2, ceil_mode=True)self.inception5a = Inception(832, 256, 160, 320, 32, 128, 128)self.inception5b = Inception(832, 384, 192, 384, 48, 128, 128)if self.aux_logits:  # 如果使用辅助分类器self.aux1 = InceptionAux(512, num_classes)   # 创建辅助分类器1self.aux2 = InceptionAux(528, num_classes)   # 创建辅助分类器2self.avgpool = nn.AdaptiveAvgPool2d((1, 1))self.dropout = nn.Dropout(0.4)  # 40%比例随机失活self.fc = nn.Linear(1024, num_classes)if init_weights:self._initialize_weights()  # 如果需要初始化，将会进入模型权重初始化函数

定义正向传播

将整个层结构输出出来。

if self.training and self.aux_logits:
aux2 = self.aux2(x)

判断模型是处于训练模式还是验证模式？并判断是否要使用辅助分类器。

    def forward(self, x):x = self.conv1(x)x = self.maxpool1(x)x = self.conv2(x)x = self.conv3(x)x = self.maxpool2(x)x = self.inception3a(x)x = self.inception3b(x)x = self.maxpool3(x)x = self.inception4a(x)if self.training and self.aux_logits:aux1 = self.aux1(x)x = self.inception4a(x)x = self.inception4b(x)x = self.inception4c(x)x = self.inception4d(x)if self.training and self.aux_logits:aux2 = self.aux2(x)x = self.inception4e(x)x = self.maxpool4(x)x = self.inception5a(x)x = self.inception5b(x)x = self.avgpool(x)x = torch.flatten(x, 1)x = self.dropout(x)x = self.fc(x)if self.training and self.aux_logits:  # 如果处于训练模式并使用了辅助分类器return x, aux2, aux1  # 将返回 主分类器、辅助分类器2、辅助分类器1 这样3个参数return x  # 否则只返回主分类器这一个函数

初始化权重函数

    def _initialize_weights(self):for m in self.modules():if isinstance(m, nn.Conv2d):  # 卷积层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.Linear):  # 全连接层nn.init.normal_(m.weight, 0, 0.01)nn.init.constant_(m.bias, 0)