【Pytorch】学习笔记（五）

卷积神经网络（加强）

0. 1*1的卷积有什么用呢？

1. 降低运算量
2. 信息融合
3. 改变通道数量
   
   

1. GoogLeNet

Inception Block

沿着通道的维度进行拼接

完整代码

"""
2021.08.10
author:alian
11 基于pytorch的卷积神经网络（GoogleNet）
"""
import torch
import torch.nn.functional as F
# GoogleNet
class InceptionA(torch.nn.Module):
    def __init__(self,in_channels):
        super(InceptionA,self).__init__()
        self.branch1x1 = torch.nn.Conv2d(in_channels,16, kernel_size=1)

        self.branch5x5_1 = torch.nn.Conv2d(in_channels, 16, kernel_size=1)
        self.branch5x5_2 = torch.nn.Conv2d(16, 24, kernel_size=5, padding=2)

        self.branch3x3_1 = torch.nn.Conv2d(in_channels, 16, kernel_size=1)
        self.branch3x3_2 = torch.nn.Conv2d(16, 24, kernel_size=3, padding=1)
        self.branch3x3_3 = torch.nn.Conv2d(24, 24, kernel_size=3, padding=1)

        self.branch_pool = torch.nn.Conv2d(in_channels, 24, kernel_size=1)

    def forward(self,x):
        branch1x1 = self.branch1x1(x)

        branch5x5 = self.branch5x5_1(x)
        branch5x5 = self.branch5x5_2(branch5x5)

        branch3x3 = self.branch3x3_1(x)
        branch3x3 = self.branch3x3_2(branch3x3)
        branch3x3 = self.branch3x3_3(branch3x3)

        branch_pool = F.avg_pool2d(x,kernel_size=3,stride=1,padding=1)
        branch_pool = self.branch_pool(branch_pool)

        outputs = [branch1x1,branch5x5,branch3x3,branch_pool]
        return torch.cat(outputs,dim=1)

class googleNet(torch.nn.Module):
    def __init__(self):
        super(googleNet,self).__init__()
        self.conv1 = torch.nn.Conv2d(1,10,kernel_size=5)
        self.conv2 = torch.nn.Conv2d(88,20,kernel_size=5)

        self.incep1 = InceptionA(in_channels=10)
        self.incep2 = InceptionA(in_channels=20)

        self.mp = torch.nn.MaxPool2d(2)
        self.fc = torch.nn.Linear(1408,10)  # 全连接层（线性层）

    def forward(self,x):
        in_size = x.size(0)
        x = F.relu(self.mp(self.conv1(x)))
        x = self.incep1(x)
        x = F.relu(self.mp(self.conv2(x)))
        x = self.incep2(x)
        x = x.view(in_size,-1)
        x = self.fc(x)
        return x

2. ResNet

梯度消失：
反向传播，链式法则求梯度（一连串地将梯度乘起来），假如每一处的梯度都是<1，则一连串地相乘会导致梯度趋近于0，当梯度趋近于0时，根据权重更新公式：$w=w-\alpha g$ ($g$为梯度)，权重将得不到更新，无法得到充分训练。

上述的残差单元，在输出加上一个输入x，将来反向传播回传的梯度（即使非常小）也趋近于1，链式法则梯度相乘不会造成梯度消失问题，开始的网络层能得到很好的训练。要求：残差单元输出和输入的张量维度必须一致

完整代码

# ResNet
class ResidualBlock(torch.nn.Module):
    def __init__(self,channels):
        super(ResidualBlock,self).__init__()
        self.channels = channels
        self.conv1 = torch.nn.Conv2d(channels,channels,kernel_size=3,padding=1)
        self.conv2 = torch.nn.Conv2d(channels, channels, kernel_size=3, padding=1)

    def forward(self,x):
        y = F.relu(self.conv1(x))
        y = self.conv2(y)
        return F.relu(x+y)

class ResNet(torch.nn.Module):
    def __init__(self):
        super(ResNet,self).__init__()
        self.conv1 = torch.nn.Conv2d(1,16,kernel_size=5)
        self.conv2 = torch.nn.Conv2d(16,32,kernel_size=5)
        self.res1 = ResidualBlock(16)
        self.res2 = ResidualBlock(32)
        self.pool = torch.nn.MaxPool2d(2)
        self.fc = torch.nn.Linear(512,10)

    def forward(self,x):  # 输入张量
        in_size = x.size(0)
        x = self.pool(F.relu(self.conv1(x)))
        x = self.res1(x)
        x = self.pool(F.relu(self.conv2(x)))
        x = self.res2(x)
        x.view(in_size,-1) # Flatten data from(batch_size,32,4,4) to (batch_size,512)
        x = self.fc(x)
        return x