pytorch--CIFAR-10

本期我们来利用pytorch深度学习框架进行CIFAR-10项目实践。从本文你将要学到

• 如何利用torchvision.datasets读取远程数据

CIFAR-10图像分类项目

• ​​背景​​
• ​​读取数据并可视化​​
• ​​构建网络，损失函数，优化方式​​
• ​​模型训练​​
• ​​评估模型​​
• ​​保存模型​​
• ​​加载模型做测试​​
• ​​参考文献​​

背景

CIFAR-10是​​kaggle计算机视觉竞赛​​的一个图像分类项目。该数据集共有60000张32*32彩色图像，一共可以分为"plane", “car”, “bird”,“cat”, “deer”, “dog”, “frog”,“horse”,“ship”, “truck” 10类，每类6000张图。有50000张用于训练，构成了5个训练批，每一批10000张图；10000张用于测试，单独构成一批。

读取数据并可视化

import torch
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")


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("%5s" % classes[labels[j]] for j in range(4)))

• 输出结果

Files already downloaded and verified

Files already downloaded and verified

truck car frog plane

构建网络，损失函数，优化方式

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

class Net(nn.Module): #继承的torch.nn.Module类
def __init__(self):
super(Net, self).__init__()
self.conv1 = nn.Conv2d(3, 6, 5) #添加第一个卷积层,调用了nn里面的Conv2d()
self.pool = nn.MaxPool2d(2, 2) #添加最大池化层
self.conv2 = nn.Conv2d(6, 16, 5) #添加第二个卷积层
self.fc1 = nn.Linear(16*5*5, 120) #第一个全连接层
self.fc2 = nn.Linear(120, 84) #第二个全连接层
self.fc3 = nn.Linear(84, 10) #第三个全连接层

def forward(self, x): #定义向前传播方法
x = self.pool(F.relu(self.conv1(x))) #relu激活第一个卷积层
x = self.pool(F.relu(self.conv2(x))) #relu激活第二个卷积层
x = x.view(-1, 16*5*5) #重构张量的维度
x = F.relu(self.fc1(x))
x = F.relu(self.fc2(x))
x = self.fc3(x)
return x

net = Net() #网络实例化
criterion = nn.CrossEntropyLoss() #定义损失函数为交叉熵损失函数
optimizer = optim.SGD(net.parameters(), lr = 0.001, momentum = 0.9) #定义优化方式为机梯度下降

模型训练

for epoch in range(2):

running_loss =0.0
for i, data in enumerate(trainloader, 0):
inputs, labels = data
optimizer.zero_grad()
outputs = net(inputs)
loss = criterion(outputs, labels)
loss.backward()
optimizer.step()

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

[1, 2000] loss: 2.221

[1, 4000] loss: 1.908

[1, 6000] loss: 1.692

[1, 8000] loss: 1.595

[1, 10000] loss: 1.536

[1, 12000] loss: 1.475

[2, 2000] loss: 1.405

[2, 4000] loss: 1.376

[2, 6000] loss: 1.345

[2, 8000] loss: 1.321

[2, 10000] loss: 1.299

[2, 12000] loss: 1.279

Finished Training

评估模型

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))

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]))

保存模型

PATH =".cifar10_net.pth"
torch.save(net.state_dict(), PATH)

加载模型做测试

net = Net()
net.load_state_dict(torch.load(PATH))

outputs = net(images)

_, predicted = torch.max(outputs, 1)
print("predicted: ", " ".join("%5s" % classes[predicted[j]] for j in range(4)))

参考文献