P2_CIFAR10彩色图片识别

• 🍨 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

第P2周：彩色图片识别
● 难度：小白入门
● 语言：Python3、Pytorch

要求：

1. 学习如何编写一个完整的深度学习程序
2. 手动推导卷积层与池化层的计算过程

本次的重点在于学会构建CNN网络：

🏡 我的环境：

● 语言环境：Python3.8
● 编译器：jupyter notebook
● 深度学习环境：Pytorch
  ○ torch == 2.2.2 + cu121
  ○ torchvision == 0.17.2 + cu121

一、前期准备

1.设置GPU

import torch
import torch.nn as nn
import matplotlib.pyplot as plt
import torchvision

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

device

device(type='cuda')

2.导入数据

train_ds = torchvision.datasets.CIFAR10('data', 
                                      train=True, 
                                      transform=torchvision.transforms.ToTensor(), # 将数据类型转化为Tensor
                                      download=True)

test_ds  = torchvision.datasets.CIFAR10('data', 
                                      train=False, 
                                      transform=torchvision.transforms.ToTensor(), # 将数据类型转化为Tensor
                                      download=True)

Files already downloaded and verified
Files already downloaded and verified

batch_size = 32

train_dl = torch.utils.data.DataLoader(train_ds,
                                      batch_size = batch_size,
                                      shuffle = True)

test_dl = torch.utils.data.DataLoader(test_ds,
                                     batch_size = batch_size)

# 取一个批次查看数据格式
# 数据的shape为：[batch_size, channel, height, weight]
# 其中batch_size为自己设定，channel，height和weight分别是图片的通道数，高度和宽度。
imgs, labels = next(iter(train_dl))
imgs.shape

torch.Size([32, 3, 32, 32])

3.数据可视化

import numpy as np

# 指定图片大小， 宽20、长5英寸
plt.figure(figsize = (20, 5))
for i, imgs in enumerate(imgs[: 20]):
    # 进行轴变换（转置）
    npimg = imgs.numpy().transpose((1, 2, 0))
    # 将整个figure分成2行10列，绘制第i+1个子图。
    plt.subplot(2, 10, i + 1)
    plt.imshow(npimg, cmap = plt.cm.binary)
    plt.axis('off')

二、构建简单的CNN网络

import torch.nn.functional as F

num_classes = 10 # 图片类别数

class Model(nn.Module):
    def __init__(self):
        super().__init__()
        # 特征提取网络
        self.conv1 = nn.Conv2d(3, 64, kernel_size=3)
        self.pool1 = nn.MaxPool2d(kernel_size=2)
        self.conv2 = nn.Conv2d(64, 64, kernel_size=3)
        self.pool2 = nn.MaxPool2d(kernel_size=2)
        self.conv3 = nn.Conv2d(64, 128, kernel_size=3)
        self.pool3 = nn.MaxPool2d(kernel_size=2)
        
        # 分类网络
        self.fc1 = nn.Linear(512, 256)
        self.fc2 = nn.Linear(256, num_classes)
    # 前向传播
    def forward(self, x):
        x = self.pool1(F.relu(self.conv1(x)))
        x = self.pool2(F.relu(self.conv2(x)))
        x = self.pool3(F.relu(self.conv3(x)))
        
        x = torch.flatten(x, start_dim = 1)
        
        x = F.relu(self.fc1(x))
        x = self.fc2(x)
        
        return x
    
    

from torchinfo import summary

model = Model().to(device) # 模型转移到GPU

summary(model)

=================================================================
Layer (type:depth-idx)                   Param #
=================================================================
Model                                    --
├─Conv2d: 1-1                            1,792
├─MaxPool2d: 1-2                         --
├─Conv2d: 1-3                            36,928
├─MaxPool2d: 1-4                         --
├─Conv2d: 1-5                            73,856
├─MaxPool2d: 1-6                         --
├─Linear: 1-7                            131,328
├─Linear: 1-8                            2,570
=================================================================
Total params: 246,474
Trainable params: 246,474
Non-trainable params: 0
=================================================================

三、训练模型

1.设置超参数

loss_fn = nn.CrossEntropyLoss() # 创建损失函数
learn_rate = 1e-2 # 学习率
opt = torch.optim.SGD(model.parameters(), lr = learn_rate)

2.编写训练函数

# 训练循环
def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset) # 训练集大小
    num_batches = len(dataloader)
    
    train_loss, train_acc = 0, 0
    
    for X, y in dataloader:
        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) # 网络输出与真实值之间的差距
        
        # 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                   

3.编写测试函数

def test (dataloader, model, loss_fn):
    size        = len(dataloader.dataset)  # 测试集的大小，一共10000张图片
    num_batches = len(dataloader)
    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

4.正式训练

epochs     = 10
train_loss = []
train_acc  = []
test_loss  = []
test_acc   = []

for epoch in range(epochs):
    model.train()
    epoch_train_acc, epoch_train_loss = train(train_dl, model, loss_fn, opt)
    
    model.eval()
    epoch_test_acc, epoch_test_loss = test(test_dl, model, loss_fn)
    
    train_acc.append(epoch_train_acc)
    train_loss.append(epoch_train_loss)
    test_acc.append(epoch_test_acc)
    test_loss.append(epoch_test_loss)
    
    template = ('Epoch:{:2d}, Train_acc:{:.1f}%, Train_loss:{:.3f}, Test_acc:{:.1f}%，Test_loss:{:.3f}')
    print(template.format(epoch+1, epoch_train_acc*100, epoch_train_loss, epoch_test_acc*100, epoch_test_loss))
print('Done')

Epoch: 1, Train_acc:14.5%, Train_loss:2.256, Test_acc:22.4%，Test_loss:2.087
Epoch: 2, Train_acc:25.1%, Train_loss:2.000, Test_acc:30.4%，Test_loss:1.894
Epoch: 3, Train_acc:34.0%, Train_loss:1.796, Test_acc:39.1%，Test_loss:1.658
Epoch: 4, Train_acc:40.4%, Train_loss:1.630, Test_acc:44.1%，Test_loss:1.543
Epoch: 5, Train_acc:44.6%, Train_loss:1.525, Test_acc:44.3%，Test_loss:1.521
Epoch: 6, Train_acc:48.1%, Train_loss:1.436, Test_acc:51.1%，Test_loss:1.375
Epoch: 7, Train_acc:51.5%, Train_loss:1.349, Test_acc:52.2%，Test_loss:1.325
Epoch: 8, Train_acc:54.6%, Train_loss:1.276, Test_acc:53.9%，Test_loss:1.272
Epoch: 9, Train_acc:56.9%, Train_loss:1.211, Test_acc:54.6%，Test_loss:1.272
Epoch:10, Train_acc:59.3%, Train_loss:1.154, Test_acc:59.4%，Test_loss:1.156
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

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.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、总结

dataset 的一部分参数

  dataset：需要读取的数据集
  batch_size：需要int, 表示每次抓取数据要抓几个观察
  shuffle：需要bool, 表示是否随机抓取, (有种类似种子的感觉, 如果是False, 无论怎么跑都是一样的抓取)
  num_wordker：需要int, 使用n个进程进行抓取数据 (windows下可能会报错)
  drop_last：需要book, 当样本量/batch_size 无法除尽时, 是否舍去"余数数据" (True-舍弃; False-保留)

conv2d 的一部分参数

  in_channels：输入图像(矩阵)的通道数.
  out_channels：输出(矩阵) 的通道数. 当out_channels = 2的时候, 会生成两个kernel然后生成两个输出矩阵.
  kernel_size：卷积核的size, 单个int 就是int * int大小; 元组()就是H * W大小的矩阵. 卷积核的值是从分布中随机获得, 会在训练过程调整
  stride：卷积核移动的步数(先横向, 再纵向) 可以是元组(sH, sW) (横向的步径, 纵向的步径); 也可以单个int,则横纵步径一致
  padding：在输入图像上下左右增加行列, 新增格子中默认填充数字0. 元组(padH, padW)说明填充多少
  dilation：空洞卷积, 就是卷积核会有间隔, 默认是没有间隔

nn.MaxPool的一部分参数

  kernel_size：池化核? 还是需要一个int或者元组
  stride：默认是kernel_size的大小
  ceil_model：True代表池化核+strid超过了输入矩阵的范围,仍然保留获得的元素