🍨 本文为🔗365天深度学习训练营中的学习记录博客
 ● 🍖 原作者：K同学啊

学习目标：

1.保存最佳模型权重

2.调用官方VGG-16网络框架

一、前期准备

1.设置GPU

import torch
import torch.nn as nn
import torchvision.transforms as transforms
import torchvision
from torchvision import transforms, datasets
import os,PIL,pathlib,warnings

warnings.filterwarnings("ignore")             #忽略警告信息

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

2.导入数据

import os,PIL,random,pathlib

data_dir = './48-data/'
data_dir = pathlib.Path(data_dir)

data_paths  = list(data_dir.glob('*'))
classeNames = [str(path).split("\\")[1] for path in data_paths]
classeNames

train_transforms = transforms.Compose([
    transforms.Resize([224, 224]),  # 将输入图片resize成统一尺寸
    transforms.RandomHorizontalFlip(), # 随机水平翻转
    transforms.ToTensor(),          # 将PIL Image或numpy.ndarray转换为tensor，并归一化到[0,1]之间
    transforms.Normalize(           # 标准化处理-->转换为标准正太分布（高斯分布），使模型更容易收敛
        mean=[0.485, 0.456, 0.406], 
        std=[0.229, 0.224, 0.225])  # 其中 mean=[0.485,0.456,0.406]与std=[0.229,0.224,0.225] 从数据集中随机抽样计算得到的。
])

total_data = datasets.ImageFolder("./48-data/",transform=train_transforms)
total_data

total_data.class_to_idx

3.划分数据集

train_size = int(0.8 * len(total_data))
test_size  = len(total_data) - train_size
train_dataset, test_dataset = torch.utils.data.random_split(total_data, [train_size, test_size])
train_dataset, test_dataset

batch_size = 32

train_dl = torch.utils.data.DataLoader(train_dataset,
                                           batch_size=batch_size,
                                           shuffle=True,
                                           num_workers=1)
test_dl = torch.utils.data.DataLoader(test_dataset,
                                          batch_size=batch_size,
                                          shuffle=True,
                                          num_workers=1)

for X, y in test_dl:
    print("Shape of X [N, C, H, W]: ", X.shape)
    print("Shape of y: ", y.shape, y.dtype)
    break

二、调用官方的VGG-16模型

VGG-16是由牛津大学几何组提出的深度卷积神经网络架构，用于图像分类和对象识别任务。VGG-16在ImageNet图像识别竞赛中取得了很好的成绩，展示了其在大规模识别任务中的有效性。

VGG-16的主要特点：

1.深度：VGG-16由13个卷积层和3个全连接层组成，因此具有相对较深的网络结构，这种深度有助于网络学习的更加抽象和复杂的特征。

2.卷积层的设计：VGG-16的卷积层全部采用3×3卷积核和步长为1的卷积操作，同时在卷积层之后都接有ReLU激活函数：ReLU(x)=max(0,x)。这种设计的好处在于，堆叠多个较小的卷积核，可以提高网络的非线性建模能力，同时减少了参数数量，从而降低了过拟合的风险。

3.池化层：在卷积核之后，VGG-16使用最大化池化层来减少特征图的空间尺寸，帮助提取更加显著的特征并减少计算量。

4.全连接层：VGG-16在卷积层之后接有3个全连接层，最后一个全连接层输出与类别相对应的向量用于进行分类。

VGG-16的网络结构：13个卷积层，3个全连接层，5个池化层

from torchvision.models import vgg16

device = "cuda" if torch.cuda.is_available() else "cpu"
print("Using {} device".format(device))
    
# 加载预训练模型，并且对模型进行微调
model = vgg16(pretrained = True).to(device) # 加载预训练的vgg16模型

for param in model.parameters():
    param.requires_grad = False # 冻结模型的参数，这样子在训练的时候只训练最后一层的参数

# 修改classifier模块的第6层（即：(6): Linear(in_features=4096, out_features=2, bias=True)）
# 注意查看我们下方打印出来的模型
model.classifier._modules['6'] = nn.Linear(4096,len(classeNames)) # 修改vgg16模型中最后一层全连接层，输出目标类别个数
model.to(device)  
model

三、训练模型

1.编写训练函数

# 训练循环
def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset)  # 训练集的大小
    num_batches = len(dataloader)   # 批次数目, (size/batch_size，向上取整)

    train_loss, train_acc = 0, 0  # 初始化训练损失和正确率
    
    for X, y in dataloader:  # 获取图片及其标签
        X, y = X.to(device), y.to(device)
        
        # 计算预测误差
        pred = model(X)          # 网络输出
        loss = loss_fn(pred, y)  # 计算网络输出和真实值之间的差距，targets为真实值，计算二者差值即为损失
        
        # 反向传播
        optimizer.zero_grad()  # grad属性归零
        loss.backward()        # 反向传播
        optimizer.step()       # 每一步自动更新
        
        # 记录acc与loss
        train_acc  += (pred.argmax(1) == y).type(torch.float).sum().item()
        train_loss += loss.item()
            
    train_acc  /= size
    train_loss /= num_batches

    return train_acc, train_loss

2.编写测试函数

def test (dataloader, model, loss_fn):
    size        = len(dataloader.dataset)  # 测试集的大小
    num_batches = len(dataloader)          # 批次数目, (size/batch_size，向上取整)
    test_loss, test_acc = 0, 0
    
    # 当不进行训练时，停止梯度更新，节省计算内存消耗
    with torch.no_grad():
        for imgs, target in dataloader:
            imgs, target = imgs.to(device), target.to(device)
            
            # 计算loss
            target_pred = model(imgs)
            loss        = loss_fn(target_pred, target)
            
            test_loss += loss.item()
            test_acc  += (target_pred.argmax(1) == target).type(torch.float).sum().item()

    test_acc  /= size
    test_loss /= num_batches

    return test_acc, test_loss

3.设置动态学习率

def adjust_learning_rate(optimizer, epoch, start_lr):     # 每 2 个epoch衰减到原来的 0.98
    lr = start_lr * (0.92 ** (epoch // 2))
    for param_group in optimizer.param_groups:
        param_group['lr'] = lr
learn_rate = 1e-4 # 初始学习率
optimizer  = torch.optim.SGD(model.parameters(), lr=learn_rate)

4.正式训练并保存模型

import copy

loss_fn    = nn.CrossEntropyLoss() # 创建损失函数
epochs     = 40

train_loss = []
train_acc  = []
test_loss  = []
test_acc   = []

best_acc = 0    # 设置一个最佳准确率，作为最佳模型的判别指标

for epoch in range(epochs):
    # 更新学习率（使用自定义学习率时使用）
    adjust_learning_rate(optimizer, epoch, learn_rate)
    
    model.train()
    epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, optimizer)
    #scheduler.step() # 更新学习率（调用官方动态学习率接口时使用）
    
    model.eval()
    epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)
    
    # 保存最佳模型到 best_model
    if epoch_test_acc > best_acc:
        best_acc   = epoch_test_acc
        best_model = copy.deepcopy(model)
    
    train_acc.append(epoch_train_acc)
    train_loss.append(epoch_train_loss)
    test_acc.append(epoch_test_acc)
    test_loss.append(epoch_test_loss)
    
    # 获取当前的学习率
    lr = optimizer.state_dict()['param_groups'][0]['lr']
    
    template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%, Test_loss:{:.3f}, Lr:{:.2E}')
    print(template.format(epoch+1, epoch_train_acc*100, epoch_train_loss, 
                          epoch_test_acc*100, epoch_test_loss, lr))
    
# 保存最佳模型到文件中
PATH = './best_model.pth'  # 保存的参数文件名
torch.save(best_model.state_dict(), PATH)

print('Done')

四、结果可视化

import matplotlib.pyplot as plt
#隐藏警告
import warnings
warnings.filterwarnings("ignore")               #忽略警告信息
plt.rcParams['font.sans-serif']    = ['SimHei'] # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False      # 用来正常显示负号
plt.rcParams['figure.dpi']         = 100        #分辨率

from datetime import datetime
current_time = datetime.now() # 获取当前时间

epochs_range = range(epochs)

plt.figure(figsize=(12, 3))
plt.subplot(1, 2, 1)

plt.plot(epochs_range, train_acc, label='Training Accuracy')
plt.plot(epochs_range, test_acc, label='Test Accuracy')
plt.legend(loc='lower right')
plt.title('Training and Validation Accuracy')
plt.xlabel(current_time) 

plt.subplot(1, 2, 2)
plt.plot(epochs_range, train_loss, label='Training Loss')
plt.plot(epochs_range, test_loss, label='Test Loss')
plt.legend(loc='upper right')
plt.title('Training and Validation Loss')
plt.show()

五、手动搭建VGG-16网络框架

import torch.nn.functional as F

class Model(nn.Module):
    def __init__(self):
        super(Model, self).__init__()
        self.conv1=nn.Sequential(
            nn.Conv2d(3, 64, kernel_size=3, padding=1), # 224*224*64
            nn.ReLU(inplace=True))
        
        self.conv2=nn.Sequential(
            nn.Conv2d(64, 64, kernel_size=3, padding=1), # 224*224*64
            nn.ReLU(inplace=True))
        
        self.pool1=nn.Sequential(
            nn.MaxPool2d(2))                              # 112*112*64
        
        self.conv3=nn.Sequential(
            nn.Conv2d(64, 128, kernel_size=3, padding=1), # 112*112*128
            nn.ReLU(inplace=True))
        
        self.conv4=nn.Sequential(
            nn.Conv2d(128, 128, kernel_size=3, padding=1), # 112*112*128
            nn.ReLU(inplace=True))
        
        self.pool2=nn.Sequential(
            nn.MaxPool2d(2))                               # 56*56*128
        
        self.conv5=nn.Sequential(
            nn.Conv2d(128, 256, kernel_size=3, padding=1), # 56*56*256
            nn.ReLU(inplace=True))

        self.conv6=nn.Sequential(
            nn.Conv2d(256, 256, kernel_size=3, padding=1), # 56*56*256
            nn.ReLU(inplace=True))

        self.conv7=nn.Sequential(
            nn.Conv2d(256, 256, kernel_size=3, padding=1), # 56*56*256
            nn.ReLU(inplace=True))

        self.pool3=nn.Sequential(
            nn.MaxPool2d(2))                               #28*28*256

        self.conv8=nn.Sequential(
            nn.Conv2d(256, 512, kernel_size=3, padding=1), #28*28*512
            nn.ReLU(inplace=True))

        self.conv9=nn.Sequential(
            nn.Conv2d(512, 512, kernel_size=3, padding=1), # 28*28*512
            nn.ReLU(inplace=True))

        self.conv10=nn.Sequential(
            nn.Conv2d(512, 512, kernel_size=3, padding=1), # 28*28*512
            nn.ReLU(inplace=True))

        self.pool4=nn.Sequential(
            nn.MaxPool2d(2))                               #14*14*512

        self.conv11=nn.Sequential(
            nn.Conv2d(512, 512, kernel_size=3, padding=1), # 14*14*512
            nn.ReLU(inplace=True))

        self.conv12=nn.Sequential(
            nn.Conv2d(512, 512, kernel_size=3, padding=1), # 14*14*512
            nn.ReLU(inplace=True))
        
        self.conv13=nn.Sequential(
            nn.Conv2d(512, 512, kernel_size=3, padding=1), # 14*14*512
            nn.ReLU(inplace=True))

        self.pool5=nn.Sequential(
            nn.MaxPool2d(2))                               #  7*7*512
        
        self.fc1=nn.Sequential(
            nn.Linear(7*7*512, 4096),
            nn.ReLU(inplace=True),
            nn.Dropout(0.5))
        
        self.fc2=nn.Sequential(
            nn.Linear(4096,4096),
            nn.ReLU(inplace=True),
            nn.Dropout(0.5))
        
        self.fc3=nn.Sequential(
            nn.Linear(4096, len(classeNames)))
        
    def forward(self, x):
        
        batch_size = x.size(0)
        x = self.conv1(x)  # 卷积-激活
        x = self.conv2(x)  # 卷积-激活
        x = self.pool1(x)  # 池化
        x = self.conv3(x)  # 卷积-激活
        x = self.conv4(x)  # 卷积-激活
        x = self.pool2(x)  # 池化
        x = self.conv5(x)  # 卷积-激活
        x = self.conv6(x)  # 卷积-激活
        x = self.conv7(x)  # 卷积-激活
        x = self.pool3(x)  # 池化
        x = self.conv8(x)  # 卷积-激活
        x = self.conv9(x)  # 卷积-激活
        x = self.conv10(x) # 卷积-激活
        x = self.pool4(x)  # 池化
        x = self.conv11(x) # 卷积-激活
        x = self.conv12(x) # 卷积-激活
        x = self.conv13(x) # 卷积-激活
        x = self.pool5(x)  # 池化
        x = x.view(batch_size, -1)
        x = self.fc1(x)
        x = self.fc2(x)
        x = self.fc3(x)
       
        return x

device = "cuda" if torch.cuda.is_available() else "cpu"
print("Using {} device".format(device))

model = Model().to(device)
model

六、模型调优

尝试提高测试集准确率，在AI的建议下

from torchvision.models import vgg16

device = "cuda" if torch.cuda.is_available() else "cpu"
print("Using {} device".format(device))
    
model = vgg16(pretrained=True).to(device)

# ----------------------核心冻结逻辑：冻结前3个卷积块，解冻后2个卷积块----------------------
# VGG16 features层结构划分：
# Block1: 0~1层  | Block2: 2~4层 | Block3:5~8层 | Block4:9~12层 | Block5:13~16层
# 冻结前3个Block(0~8层)，解冻Block4、Block5(9~16层)
for idx, param in enumerate(model.features.parameters()):
    if idx < 17:  # 前17个参数对应前3个卷积块，完全冻结
        param.requires_grad = False
    else:         # 后两个卷积块Block4、Block5，允许梯度更新微调人脸高层特征
        param.requires_grad = True
# ----------------------------------------------------------------------------------------

# 2. 替换最后一层全连接，适配你的人脸分类数
num_classes = len(classeNames)
in_features = model.classifier[6].in_features
model.classifier[6] = nn.Linear(in_features, num_classes)


model.to(device)

冻结前3个卷积块

浅层（前 3 个卷积块）：负责提取通用基础特征（横竖边、明暗梯度、基础纹理），人脸、猫狗、风景都共用这套基础视觉逻辑，完全不需要重新训练；

冻结操作：这部分网络权重锁死不更新，反向传播时不会修改参数

解冻范围：VGG16 后 2 个卷积块（block4、block5）+ 最后的全连接分类头

个人总结：本周学习了VGG-16算法，通过代码复现了VGG-16的框架，这是目前学到的层数最多的模型，有13个卷积层和3个全连接层，每个卷积层后面都有激活函数。提高训练集准确率时，我使用了AI给定的代码，使VGG-16网络的前3个卷积块冻结，后面解冻，使训练集的准确率提高到60%以上，不过模型过拟合很严重。