基于Pytorch的深度学习-第二章

2.1 CIFAR-10数据集简介

CIFAR-10数据集包含10个类别：plane、car、bird、cat、deer、dog、frog、horse、ship、truck，每个类别有6000张图片。其中训练集图片有50000张，测试集有10000张图片。训练集和测试集的生成方法是，分别从每个类别中随机挑选1000张图片加入测试集，其余图片进入训练集。CIFAR-10中的图像尺寸为33232，也就是RGB的3层颜色通道，图像的宽和高都为32。

2.2 加载数据集

加载数据集，数据进行归一化操作

import torch
import torch.utils
import torchvision
import torchvision.transforms as transforms

transform = transforms.Compose(
    [transforms.ToTensor(),
     transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]
)

trainset = torchvision.datasets.CIFAR10(root='./data', train=True,
                                        download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=4,
                                          shuffle=True, num_workers=0)
testset = torchvision.datasets.CIFAR10(root='./data', train=False,download=True,transform=transform)

testloader = torch.utils.data.DataLoader(testset, batch_size=4,shuffle=False, num_workers=0)

classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck')

展示一些训练图片，代码如下：

import matplotlib.pyplot as plt
import numpy as np

def imshow(img):
    img = img / 2 + 0.5
    npimg = img.numpy()
    plt.imshow(np.transpose(npimg, (1,2,0)))
    plt.show()
dataiter = iter(trainloader)
images, labels = dataiter.next()

imshow(torchvision.utils.make_grid(images))
print(''.join('%6s' % classes[labels[j]] for j in range(4)))

输出结果：

  deer plane  deer   cat

2.3定义卷积神经网络

import torch.nn as nn
import torch.nn.functional as F

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        #输入图片为三通道，输出为六通道，卷积核大小为5*5
        self.conv1 = nn.Conv2d(3, 6, 5)
        self.pool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(6, 16, 5)
        self.fc1 = nn.Linear(16 * 5 * 5, 120)
        self.fc2 = nn.Linear(120, 84)
        self.fc3 = nn.Linear(84, 10)
    def forward(self, x):
        x = self.pool(F.relu(self.conv1(x)))
        x = self.pool(F.relu(self.conv2(x)))
        x = x.view(-1, 16 * 5 * 5)
        x = F.relu(self.fc1(x))
        x = F.relu(self.fc2(x))
        return x

net = Net()
print(net)

输出网络结构：

Net(
  (conv1): Conv2d(3, 6, kernel_size=(5, 5), stride=(1, 1))
  (pool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv2d(6, 16, kernel_size=(5, 5), stride=(1, 1))
  (fc1): Linear(in_features=400, out_features=120, bias=True)
  (fc2): Linear(in_features=120, out_features=84, bias=True)
  (fc3): Linear(in_features=84, out_features=10, bias=True)
)

2.4 定义损失函数和优化器

使用分类交叉熵（CrossEntropy）作为损失函数，支持动量的SGD作为优化器，代码如下：

import torch.optim as optim

criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(net.parameters(), lr=0.001, momentum=0.9)

动量设置为0.9。

2.5 训练网络

只需要在数据迭代器上循环传给网络和优化器的输入即可，代码如下：

for epoch in range(2):
    running_loss = 0.0
    for i, data in enumerate(trainloader, 0):
        #获取输入
        inputs, labels = data

        #清零权重的梯度
        optimizer.zero_grad()
        #前向传播 计算损失 反向传播 更新参数
        outputs = net(inputs)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()

        #打印统计信息
        running_loss += loss.item()
        if i % 2000 == 1999:
            print('[%d, %5d] loss: %.3f' %
                  (epoch + 1, i + 1,running_loss / 2000))
            running_loss = 0.0

print('Finished Training')

输出结果如下：

[1,  2000] loss: 2.695
[1,  4000] loss: 2.166
[1,  6000] loss: 2.026
[1,  8000] loss: 1.646
[1, 10000] loss: 1.483
[1, 12000] loss: 1.440
[2,  2000] loss: 1.327
[2,  4000] loss: 1.345
[2,  6000] loss: 1.326
[2,  8000] loss: 1.295
[2, 10000] loss: 1.239
[2, 12000] loss: 1.232
Finished Training

2.6使用测试集评估

从测试集选取一些图片，来用训练好的网络来进行预测，代码如下：

dataiter = iter(testloader)
images, labels = dataiter.next()
imshow(torchvision.utils.make_grid(images))
print('GroundTruth', ''.join('%6s' % classes[labels[j]] for j in range(4)))

outputs = net(images)
_, predicted = torch.max(outputs, 1)
print('Predicted: ',' '.join('%6s' % classes[predicted[j]] for j in range(4)))

输出结果如下：

GroundTruth    cat  ship  ship plane
Predicted:     cat   ship   ship   ship

接下来对测试集的每一张图片都进行预测，并计算整体的准确率，代码如下：

correct = 0
total = 0
with torch.no_grad():
    for data in testloader:
        images, labels =  data
        outputs = net(images)
        _, predicted = torch.max(outputs.data, 1)
        total += labels.size(0)
        correct += (predicted == labels).sum().item()

print('Accuracy of the network on the 10000 test images: %d %%' % (100 * correct / total))

输出结果如下：

Accuracy of the network on the 10000 test images: 56 %

为了精细化分析，查看每一个类别的准确率。代码如下：

class_correct = list(0. for i in range(10))
class_total = list(0. for i in range(10))
with torch.no_grad():
    for data in testloader:
        images, labels = data
        outputs = net(images)
        _, predicted = torch.max(outputs, 1)
        c = (predicted == labels).squeeze()
        for i in range(4):
            label = labels[i]
            class_correct[label] += c[i].item()
            class_total[label] += 1

for i in range(10):
    print('Accuracy of %5s : %2d %%' % (classes[i], 100 * class_correct[i] / class_total[i]))

输出结果如下：

Accuracy of plane : 41 %
Accuracy of   car : 81 %
Accuracy of  bird : 27 %
Accuracy of   cat : 41 %
Accuracy of  deer : 61 %
Accuracy of   dog : 46 %
Accuracy of  frog : 73 %
Accuracy of horse : 59 %
Accuracy of  ship : 75 %
Accuracy of truck : 58 %

如果想提高准确率，可以多训练几个epoch。