基于Pytorch构建DenseNet网络对cifar-10进行分类

news2024/10/5 19:12:25

DenseNet是指Densely connected convolutional networks（密集卷积网络）。它的优点主要包括有效缓解梯度消失、特征传递更加有效、计算量更小、参数量更小、性能比ResNet更好。它的缺点主要是较大的内存占用。

DenseNet网络与Resnet、GoogleNet类似，都是为了解决深层网络梯度消失问题的网络。

Resnet从深度方向出发，通过建立前面层与后面层之间的“短路连接”或“捷径”，从而能训练出更深的CNN网络。

GoogleNet从宽度方向出发，通过Inception(利用不同大小的卷积核实现不同尺度的感知，最后进行融合来得到图像更好的表征)。

DenseNet从特征入手，通过对前面所有层与后面层的密集连接，来极致利用训练过程中的所有特征，进而达到更好的效果和减少参数。

DenseNet网络

Dense Block：像GoogLeNet网络由Inception模块组成、ResNet网络由残差块(Residual Building Block)组成一样，DenseNet网络由Dense Block组成，论文截图如下所示：每个层从前面的所有层获得额外的输入，并将自己的特征映射传递到后续的所有层，使用级联(Concatenation)方式，每一层都在接受来自前几层的”集体知识(collective knowledge)”。增长率(growth rate)k是每个层的额外通道数。

其实说了那么多我也不大明白原理和数学推理，只需要按照相关代码做就行了

class Bottleneck(nn.Module):
    def __init__(self, input_channel, growth_rate):
        super(Bottleneck, self).__init__()
        self.bn1 = nn.BatchNorm2d(input_channel)
        self.relu1 = nn.ReLU(inplace=True)
        self.conv1 = nn.Conv2d(input_channel, 4 * growth_rate, kernel_size=1)
        self.bn2 = nn.BatchNorm2d(4 * growth_rate)
        self.relu2 = nn.ReLU(inplace=True)
        self.conv2 = nn.Conv2d(4 * growth_rate, growth_rate, kernel_size=3, padding=1)
    def forward(self, x):
        out = self.conv1(self.relu1(self.bn1(x)))
        out = self.conv2(self.relu2(self.bn2(out)))
        out = torch.cat([out, x], 1)
        return out
class Transition(nn.Module):
    def __init__(self, input_channels, out_channels):
        super(Transition, self).__init__()
        self.bn = nn.BatchNorm2d(input_channels)
        self.relu = nn.ReLU(inplace=True)
        self.conv = nn.Conv2d(input_channels, out_channels, kernel_size=1)
    def forward(self, x):
        out = self.conv(self.relu(self.bn(x)))
        out = F.avg_pool2d(out, 2)
        return out
class DenseNet(nn.Module):
    def __init__(self, nblocks, growth_rate, reduction, num_classes):
        super(DenseNet, self).__init__()
        self.growth_rate = growth_rate
        num_planes = 2 * growth_rate
        self.basic_conv = nn.Sequential(
            nn.Conv2d(3, 2 * growth_rate, kernel_size=7, stride=2, padding=3),
            nn.BatchNorm2d(2 * growth_rate),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2, padding=1)
        )
        self.dense1 = self._make_dense_layers(num_planes, nblocks[0])
        num_planes += nblocks[0] * growth_rate
        out_planes = int(math.floor(num_planes * reduction))
        self.trans1 = Transition(num_planes, out_planes)
        num_planes = out_planes
        self.dense2 = self._make_dense_layers(num_planes, nblocks[1])
        num_planes += nblocks[1] * growth_rate
        out_planes = int(math.floor(num_planes * reduction))
        self.trans2 = Transition(num_planes, out_planes)
        num_planes = out_planes
        self.dense3 = self._make_dense_layers(num_planes, nblocks[2])
        num_planes += nblocks[2] * growth_rate
        out_planes = int(math.floor(num_planes * reduction))
        self.trans3 = Transition(num_planes, out_planes)
        num_planes = out_planes
        self.dense4 = self._make_dense_layers(num_planes, nblocks[3])
        num_planes += nblocks[3] * growth_rate
        self.AdaptiveAvgPool2d = nn.AdaptiveAvgPool2d(1)
        # 全连接层
        self.fc = nn.Sequential(
            nn.Linear(num_planes, 256),
            nn.ReLU(inplace=True),
            # 使一半的神经元不起作用，防止参数量过大导致过拟合
            nn.Dropout(0.5),
            nn.Linear(256, 128),
            nn.ReLU(inplace=True),
            nn.Dropout(0.5),
            nn.Linear(128, 10)
        )
    def _make_dense_layers(self, in_planes, nblock):
        layers = []
        for i in range(nblock):
            layers.append(Bottleneck(in_planes, self.growth_rate))
            in_planes += self.growth_rate
        return nn.Sequential(*layers)
    def forward(self, x):
        out = self.basic_conv(x)
        out = self.trans1(self.dense1(out))
        out = self.trans2(self.dense2(out))
        out = self.trans3(self.dense3(out))
        out = self.dense4(out)
        out = self.AdaptiveAvgPool2d(out)
        out = out.view(out.size(0), -1)
        out = self.fc(out)
        return out
def DenseNet121():
    return DenseNet([6, 12, 24, 16], growth_rate=32, reduction=0.5, num_classes=10)
def DenseNet169():
    return DenseNet([6, 12, 32, 32], growth_rate=32, reduction=0.5, num_classes=10)
def DenseNet201():
    return DenseNet([6, 12, 48, 32], growth_rate=32, reduction=0.5, num_classes=10)
def DenseNet265():
    return DenseNet([6, 12, 64, 48], growth_rate=32, reduction=0.5, num_classes=10)
# 初始化模型
from torchstat import stat
# 定义模型输出模式，GPU和CPU均可
model = DenseNet121().to(DEVICE)