【打卡】002 p2 CIFAR10彩色图片识别

打卡～

555 我的环境：
● 语言环境：Python
● 编译器：jupyter notebook
● 深度学习环境：Pytorch

>- **🍨 本文为[🔗365天深度学习训练营](https://mp.weixin.qq.com/s/0dvHCaOoFnW8SCp3JpzKxg) 中的学习记录博客**
>- **🍖 原作者：[K同学啊](https://mtyjkh.blog.csdn.net/)**

# 一、前期准备
# 1.1 设置环境

import torch
import torch.nn as nn
import matplotlib.pyplot as plt
import torchvision
device = torch.device("cpu")

 # 1.2 导入数据
# 使用dataset下载CIFAR10数据集，并划分好训练集与测试集
# 使用dataloader加载数据，并设置好基本的batch_size

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)

Downloading https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz to data/cifar-10-python.tar.gz

100%|████████████████████████| 170498071/170498071 [00:34<00:00, 4977040.72it/s]

Extracting data/cifar-10-python.tar.gz to data
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)

imgs, labels = next(iter(train_dl))
imgs.shape

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

# 1.3 数据可视化

import numpy as np

# 指定图片大小，图像大小为20宽、5高的绘图(单位为英寸inch)
plt.figure(figsize=(20, 5)) 
for i, imgs in enumerate(imgs[:20]):
    # 进行轴变换
    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')
    
plt.show()

# 二、简易cnn网络

# 2.1 构建模型

import torch.nn.functional as F

num_classes = 10  # 图片的类别数

class Model(nn.Module):
     def __init__(self):
        super().__init__()
         # 特征提取网络
        self.conv1 = nn.Conv2d(3, 64, kernel_size=3)   # 第一层卷积,输入通道为 3，输出通道为 64，卷积核大小为3*3
        self.pool1 = nn.MaxPool2d(kernel_size=2)       # 设置池化层，池化核大小为2*2
        self.conv2 = nn.Conv2d(64, 64, kernel_size=3)  # 第二层卷积,输入通道为 64，输出通道为 64，卷积核大小为3*3   
        self.pool2 = nn.MaxPool2d(kernel_size=2) 
        self.conv3 = nn.Conv2d(64, 128, kernel_size=3) # 第三层卷积,输入通道为 64，输出通道为 128，卷积核大小为3*3   
        self.pool3 = nn.MaxPool2d(kernel_size=2) 
                                      
        # 分类网络
        self.fc1 = nn.Linear(512, 256)          # 是第一个全连接层，输入维度为 512，输出维度为 256。
        self.fc2 = nn.Linear(256, num_classes)  # 是第二个全连接层，输入维度为 256，输出维度为 num_classes，即图片的类别数。
     # 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

# 在 __init__ 方法中，定义了网络的各层结构，包括特征提取部分的卷积层 conv1、conv2、conv3 和池化层 pool1、pool2、pool3，以及分类部分的全连接层 fc1 和 fc2。
# 在 forward 方法中，定义了前向传播的计算过程。输入图片经过卷积和池化操作后，被展平为一维向量，然后经过两个全连接层进行分类预测。

# 2.2 加载并打印

from torchinfo import summary
model = Model().to(device)
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
# 3.1 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)

# 3.2 编写训练函数

# 训练循环
def train(dataloader, model, loss_fn, optimizer):
    size = len(dataloader.dataset)  # 训练集的大小，一共60000张图片
    num_batches = len(dataloader)   # 批次数目，1875（60000/32）

    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

# 3.3 编写测试函数

# 测试函数和训练函数大致相同，但是由于不进行梯度下降对网络权重进行更新，所以不需要传入优化器

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)          # 批次数目，313（10000/32=312.5，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

# 3.4 正式训练

epochs     = 30
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:13.0%, Train_loss:2.284, Test_acc:17.9%，Test_loss:2.185
Epoch: 2, Train_acc:24.3%, Train_loss:2.030, Test_acc:26.7%，Test_loss:1.969
Epoch: 3, Train_acc:33.0%, Train_loss:1.821, Test_acc:38.0%，Test_loss:1.702
Epoch: 4, Train_acc:39.4%, Train_loss:1.661, Test_acc:41.1%，Test_loss:1.640
Epoch: 5, Train_acc:43.5%, Train_loss:1.550, Test_acc:44.1%，Test_loss:1.543
Epoch: 6, Train_acc:47.1%, Train_loss:1.460, Test_acc:49.7%，Test_loss:1.389
Epoch: 7, Train_acc:50.3%, Train_loss:1.379, Test_acc:51.6%，Test_loss:1.351
Epoch: 8, Train_acc:53.3%, Train_loss:1.303, Test_acc:52.0%，Test_loss:1.392
Epoch: 9, Train_acc:55.9%, Train_loss:1.239, Test_acc:56.3%，Test_loss:1.218
Epoch:10, Train_acc:58.2%, Train_loss:1.183, Test_acc:54.8%，Test_loss:1.279
Epoch:11, Train_acc:60.3%, Train_loss:1.128, Test_acc:57.8%，Test_loss:1.182
Epoch:12, Train_acc:62.1%, Train_loss:1.081, Test_acc:57.4%，Test_loss:1.247
Epoch:13, Train_acc:64.1%, Train_loss:1.035, Test_acc:61.6%，Test_loss:1.126
Epoch:14, Train_acc:65.3%, Train_loss:0.996, Test_acc:62.8%，Test_loss:1.084
Epoch:15, Train_acc:66.9%, Train_loss:0.954, Test_acc:63.3%，Test_loss:1.064
Epoch:16, Train_acc:68.3%, Train_loss:0.917, Test_acc:64.4%，Test_loss:1.030
Epoch:17, Train_acc:69.3%, Train_loss:0.883, Test_acc:64.0%，Test_loss:1.032
Epoch:18, Train_acc:70.6%, Train_loss:0.848, Test_acc:64.8%，Test_loss:1.013
Epoch:19, Train_acc:71.8%, Train_loss:0.817, Test_acc:65.0%，Test_loss:1.025
Epoch:20, Train_acc:72.8%, Train_loss:0.786, Test_acc:67.1%，Test_loss:0.943
Epoch:21, Train_acc:73.8%, Train_loss:0.754, Test_acc:67.1%，Test_loss:0.990
Epoch:22, Train_acc:74.7%, Train_loss:0.726, Test_acc:69.5%，Test_loss:0.891
Epoch:23, Train_acc:75.8%, Train_loss:0.697, Test_acc:68.2%，Test_loss:0.927
Epoch:24, Train_acc:76.8%, Train_loss:0.669, Test_acc:69.1%，Test_loss:0.925
Epoch:25, Train_acc:77.8%, Train_loss:0.642, Test_acc:69.4%，Test_loss:0.904
Epoch:26, Train_acc:78.6%, Train_loss:0.614, Test_acc:69.7%，Test_loss:0.916
Epoch:27, Train_acc:79.5%, Train_loss:0.587, Test_acc:69.3%，Test_loss:0.936
Epoch:28, Train_acc:80.6%, Train_loss:0.561, Test_acc:69.7%，Test_loss:0.932
Epoch:29, Train_acc:81.4%, Train_loss:0.535, Test_acc:69.8%，Test_loss:0.926
Epoch:30, Train_acc:82.3%, Train_loss:0.508, Test_acc:70.0%，Test_loss:0.972
Done

plt.rcParams['font.sans-serif'] = ['SimHei', 'path/to/SimHei.ttf']

# 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()