pytorch实现经典神经网络：VGG16模型之初探

文章链接
https://blog.csdn.net/weixin_44791964/article/details/102585038?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522169675238616800211588158%2522%252C%2522scm%2522%253A%252220140713.130102334.pc%255Fblog.%2522%257D&request_id=169675238616800211588158&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2_blogfirst_rank_ecpm_v1~rank_v31_ecpm-2-102585038-null-null.nonecase&utm_term=VGG&spm=1018.2226.3001.4450

VGG16原理链接
https://zhuanlan.zhihu.com/p/460777014

代码参考：
https://blog.csdn.net/m0_50127633/article/details/117045008?ops_request_misc=&request_id=&biz_id=102&utm_term=pytorch%20vgg16&utm_medium=distribute.pc_search_result.none-task-blog-2_allsobaiduweb~default-1-117045008.142^v95insert_down28v1&spm=1018.2226.3001.4187

https://blog.csdn.net/weixin_46676835/article/details/128730174?ops_request_misc=%257B%2522request%255Fid%2522%253A%2522169681442316800215096882%2522%252C%2522scm%2522%253A%252220140713.130102334…%2522%257D&request_id=169681442316800215096882&biz_id=0&utm_medium=distribute.pc_search_result.none-task-blog-2_allsobaiduend~default-5-128730174-null-null.142^v95insert_down28v1&utm_term=pytorch%20vgg16&spm=1018.2226.3001.4187
在这里插入图片描述

分解一下：
卷积提取特征，池化压缩。
1、一张原始图片被resize到(224,224,3)。
2、conv1两次[3,3]卷积网络，输出的特征层为64，输出为(224,224,64)，再2X2最大池化，输出net为(112,112,64)。

注意池化不会改变通道数

在这里插入图片描述
根据公式第一步取stride=1，padding=1

stride=2 padding=0（不用写）

 nn.Conv2d(3,64,3,1,1),
 nn.Conv2d(64,64,3,1,1)
 nn.Conv2d(3,64,3,1,1),
 nn.Conv2d(64,64,3,1,1),
 nn.MaxPool2d(2,2)

3、conv2两次[3,3]卷积网络，第一次输入的特征层为64，输出net为(112,112,128)，再2X2最大池化，输出net为(56,56,128)。

  nn.Conv2d(64,128,3,1,1),
   nn.Conv2d(128, 128, 3, 1, 1),
   nn.MaxPool2d(2,2)，

4、conv3三次[3,3]卷积网络，输入的特征层为256，输出net为(56,56,256)，再2X2最大池化，输出net为(28,28,256)。

 nn.Conv2d(128,256,3,1,1),
            nn.Conv2d(256,256,3,1,1),
            nn.Conv2d(256,256,3,1,1),
            nn.MaxPool2d(2,2)

5、conv3三次[3,3]卷积网络，输入的特征层为256，输出net为(28,28,512)，再2X2最大池化，输出net为(14,14,512)。

nn.Conv2d(256,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.MaxPool2d(2,2),

6、conv3三次[3,3]卷积网络，输出的特征层为256，输出net为(14,14,512)，再2X2最大池化，输出net为(7,7,512)。

 nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.MaxPool2d(2,2)

7、利用卷积的方式模拟全连接层，效果等同，输出net为(1,1,4096)。共进行两次。

 nn.Linear(25088,4096), #7×7×512
            nn.Linear(4096,4096),
            nn.Linear(4096,1000)

8、利用卷积的方式模拟全连接层，效果等同，输出net为(1,1,1000)。
最后输出的就是每个类的预测。

  nn.Linear(4096,1000)

初步代码框架如下：

import torch
from torch import nn

class VGG(nn.Module):
    def __init__(self):
        super(VGG, self).__init__()
        self.MyVgg=nn.Sequential(

            nn.Conv2d(3,64,3,1,1),
            nn.Conv2d(64,64,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(64,128,3,1,1),
            nn.Conv2d(128, 128, 3, 1, 1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(128,256,3,1,1),
            nn.Conv2d(256,256,3,1,1),
            nn.Conv2d(256,256,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(256,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Linear(25088,4096), #7×7×512
            nn.Linear(4096,4096),

            nn.Linear(4096,1000)

            )

补充与完善

1、记得进行数据拉平：
在这里插入图片描述
模型部分

class VGG(nn.Module):
    def __init__(self):
        super(VGG, self).__init__()
        self.MyVgg=nn.Sequential(

            nn.Conv2d(3,64,3,1,1),
            nn.Conv2d(64,64,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(64,128,3,1,1),
            nn.Conv2d(128, 128, 3, 1, 1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(128,256,3,1,1),
            nn.Conv2d(256,256,3,1,1),
            nn.Conv2d(256,256,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(256,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Flatten(),
            nn.Linear(25088,4096), #7×7×512
            nn.Linear(4096,4096),

            nn.Linear(4096,1000)

            )
    def forward(self,x):
        x=self.MyVgg(x)
        return x

2、导入数据
使用cifar10数据集

import torch
from torch import nn
import torchvision
from torch.utils.data import DataLoader

datasets_train=torchvision.datasets.CIFAR10("./data",train=True,download=True,
                                      transform=torchvision.transforms.ToTensor())
datasets_test=torchvision.datasets.CIFAR10("./data",train=False,download=True,
                                      transform=torchvision.transforms.ToTensor())
dataloader_train=DataLoader(datasets_train,batch_size=64)
dataloader_test=DataLoader(datasets_test,batch_size=64)

3、创建损失函数
使用交叉熵CrossEntropyLoss

from torch import nn
los_fun=nn.CrossEntropyLoss()

4、创建优化器:

learning_rate=0.001
optimizer=torch.optim.SGD(fenlei.parameters(),lr=learning_rate)

5、编写训练代码：

total_train_step=0
tatal_test_step=0
epoch=10


for i in range(epoch):
    print("--------第{}轮训练开始-----".format(i+1))

    for data in dataloader_train:
        imgs,targets=data
        outputs=fenlei(imgs)
        loss=los_fun(outputs,targets)

        optimizer.zero_grad()
        loss.backward()

        total_train_step=total_train_step+1
        print("训练次数：{},Loss:{}".format(total_train_step, loss))

训练时报错：
在这里插入图片描述
原因我们假设输入的是224×224×3了
然而cfar10的数据集是32×32×3
所以我们需要加入resize操作：

from torchvision import transforms
datasets_train=torchvision.datasets.CIFAR10("./data",train=True,download=True,
                                      transform=transforms.Compose([
                                                transforms.ToTensor(),
                                                transforms.Resize([224,224])
                                      ])       )

6、使用GPU训练：

#定义训练设备
device=torch.device('cuda'if torch.cuda.is_available()else 'cpu')
fenlei.to(device)
los_fun=los_fun.to(device)

我的

import torch
from torch import nn
import torchvision
from torch.utils.data import DataLoader
from torchvision import transforms
datasets_train=torchvision.datasets.CIFAR10("./data",train=True,download=True,
                                      transform=transforms.Compose([
                                                transforms.ToTensor(),
                                                transforms.Resize([224,224])
                                      ])       )
datasets_test=torchvision.datasets.CIFAR10("./data",train=False,download=True,
                                      transform=torchvision.transforms.ToTensor())
dataloader_train=DataLoader(datasets_train,batch_size=24,drop_last=True)
dataloader_test=DataLoader(datasets_test,batch_size=64,drop_last=True)

img,target=datasets_train[0]
print(img.shape)

class VGG(nn.Module):
    def __init__(self):
        super(VGG, self).__init__()
        self.MyVgg=nn.Sequential(

            nn.Conv2d(3,64,3,1,1),
            nn.Conv2d(64,64,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(64,128,3,1,1),
            nn.Conv2d(128, 128, 3, 1, 1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(128,256,3,1,1),
            nn.Conv2d(256,256,3,1,1),
            nn.Conv2d(256,256,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(256,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.Conv2d(512,512,3,1,1),
            nn.MaxPool2d(2,2),

            nn.Flatten(),
            nn.Linear(25088,4096), #7×7×512
            nn.Linear(4096,4096),
            nn.Linear(4096,1000)

            )
    def forward(self,x):
        x=self.MyVgg(x)
        return x

fenlei=VGG()


from torch import nn
los_fun=nn.CrossEntropyLoss()

learning_rate=0.001
optimizer=torch.optim.SGD(fenlei.parameters(),lr=learning_rate)

total_train_step=0
tatal_test_step=0
epoch=10

#定义训练设备
device=torch.device('cuda'if torch.cuda.is_available()else 'cpu')
fenlei.to(device)
los_fun=los_fun.to(device)



for i in range(epoch):
    print("--------第{}轮训练开始-----".format(i+1))

    for data in dataloader_train:
        imgs,targets=data
        imgs=imgs.to(device)
        targets=targets.to(device)
        outputs = fenlei(imgs)
        loss=los_fun(outputs,targets)

        optimizer.zero_grad()
        loss.backward()

        total_train_step=total_train_step+1
        if total_train_step%10==0:
            print("训练次数：{},Loss:{}".format(total_train_step, loss))

为啥人家的代码训练这么快？

在这里插入图片描述

在PyTorch中，nn.Dropout和nn.ReLU是常用的神经网络模块，分别用于正则化和激活函数。
nn.Dropout是一种正则化技术，旨在减少神经网络的过拟合问题。过拟合是指模型在训练集上表现很好，但在测试集上表现较差的现象。Dropout通过在训练过程中随机将一定比例的神经元置为0，以强制网络学习到冗余特征，从而提高模型的泛化能力。这可以防止过拟合，并提高网络的鲁棒性。
nn.ReLU是一种常用的激活函数，它被广泛应用在神经网络中。ReLU的全称是Rectified Linear Unit，它的定义很简单：对于输入x，当x小于0时，输出为0；当x大于等于0时，输出为x。ReLU函数的优点是计算简单、非线性、减轻梯度消失等。
在神经网络中，ReLU函数能够引入非线性，增加模型的拟合能力，并且减少梯度消失问题。当输入为负时，ReLU将输出为0，这有助于稀疏表示，从而使得网络更加有效地学习特征。
综上所述，nn.Dropout用于减少过拟合，提高泛化能力，而nn.ReLU用于引入非线性和解决梯度消失问题。它们在神经网络中的应用非常常见，并且经过广泛验证的有效技术。

明天看录播有助于理解他的代码:

import torch
import torchvision
import torchvision.transforms as transforms
import torch.nn as nn
import torch.optim as optim
import torch.nn.functional as F
import numpy as np
import matplotlib.pyplot as plt

transform_train = transforms.Compose(
    [transforms.Pad(4),
     transforms.ToTensor(),
     transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),
     transforms.RandomHorizontalFlip(),
     transforms.RandomGrayscale(),
     transforms.RandomCrop(32, padding=4),
     ])

transform_test = transforms.Compose(
    [
        transforms.ToTensor(),
        transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225))]
)

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=True, transform=transform_train)
trainLoader = torch.utils.data.DataLoader(trainset, batch_size=24, shuffle=True)

testset = torchvision.datasets.CIFAR10(root='./data', train=False, download=True, transform=transform_test)
testLoader = torch.utils.data.DataLoader(testset, batch_size=24, shuffle=False)

vgg = [96, 96, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512, 'M']


class VGG(nn.Module):
    def __init__(self, vgg):
        super(VGG, self).__init__()
        self.features = self._make_layers(vgg)
        self.dense = nn.Sequential(
            nn.Linear(512, 4096),
            nn.ReLU(inplace=True),
            nn.Dropout(0.4),
            nn.Linear(4096, 4096),
            nn.ReLU(inplace=True),
            nn.Dropout(0.4),
        )
        self.classifier = nn.Linear(4096, 10)

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

    def _make_layers(self, vgg):
        layers = []
        in_channels = 3
        for x in vgg:
            if x == 'M':
                layers += [nn.MaxPool2d(kernel_size=2, stride=2)]
            else:
                layers += [nn.Conv2d(in_channels, x, kernel_size=3, padding=1),
                           nn.BatchNorm2d(x),
                           nn.ReLU(inplace=True)]
                in_channels = x

        layers += [nn.AvgPool2d(kernel_size=1, stride=1)]
        return nn.Sequential(*layers)


model = VGG(vgg)
# model.load_state_dict(torch.load('CIFAR-model/VGG16.pth'))
optimizer = optim.SGD(model.parameters(), lr=0.01, weight_decay=5e-3)
loss_func = nn.CrossEntropyLoss()
scheduler = optim.lr_scheduler.StepLR(optimizer, step_size=5, gamma=0.4, last_epoch=-1)

total_times = 40
total = 0
accuracy_rate = []


def test():
    model.eval()
    correct = 0  # 预测正确的图片数
    total = 0  # 总共的图片数
    with torch.no_grad():
        for data in testLoader:
            images, labels = data
            images = images.to(device)
            outputs = model(images).to(device)
            outputs = outputs.cpu()
            outputarr = outputs.numpy()
            _, predicted = torch.max(outputs, 1)
            total += labels.size(0)
            correct += (predicted == labels).sum()
    accuracy = 100 * correct / total
    accuracy_rate.append(accuracy)
    print(f'准确率为:{accuracy}%'.format(accuracy))


for epoch in range(total_times):
    model.train()
    model.to(device)
    running_loss = 0.0
    total_correct = 0
    total_trainset = 0

    for i, (data, labels) in enumerate(trainLoader, 0):
        data = data.to(device)
        outputs = model(data).to(device)
        labels = labels.to(device)
        loss = loss_func(outputs, labels).to(device)
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        running_loss += loss.item()
        _, pred = outputs.max(1)
        correct = (pred == labels).sum().item()
        total_correct += correct
        total_trainset += data.shape[0]
        if i % 1000 == 0 and i > 0:
            print(f"正在进行第{i}次训练, running_loss={running_loss}".format(i, running_loss))
            running_loss = 0.0
    test()
    scheduler.step()

# torch.save(model.state_dict(), 'CIFAR-model/VGG16.pth')
accuracy_rate = np.array(accuracy_rate)
times = np.linspace(1, total_times, total_times)
plt.xlabel('times')
plt.ylabel('accuracy rate')
plt.plot(times, accuracy_rate)
plt.show()

print(accuracy_rate)