3 Pytorch实例——基于Logistic 回归的手写体数字识别

import torch
import torch.nn as nn
import torchvision
import torchvision.transforms as transforms


# 设置超参数 
input_size = 784   #手写体数据集每张图像分辨率为 28*28=784
num_classes = 10   #数据集共10个类别，表示数字 0~9
num_epochs = 50
batch_size = 100
learning_rate = 0.001

# 下载手写体数据集 MNIST 
train_dataset = torchvision.datasets.MNIST(root='data', 
                                           train=True, 
                                           transform=transforms.ToTensor(),
                                           download=True)

test_dataset = torchvision.datasets.MNIST(root='data', 
                                          train=False, 
                                          transform=transforms.ToTensor())

# 设置数据加载器
train_loader = torch.utils.data.DataLoader(dataset=train_dataset, 
                                           batch_size=batch_size, 
                                           shuffle=True)

test_loader = torch.utils.data.DataLoader(dataset=test_dataset, 
                                          batch_size=batch_size, 
                                          shuffle=False)

# Logistic 回归模型
model = nn.Linear(input_size, num_classes)

# 定义损失函数Loss 和优化器 optimizer
criterion = nn.CrossEntropyLoss()   #使用交叉熵作为损失函数
optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate)  

# 训练模型
total_step = len(train_loader)
for epoch in range(num_epochs):
    for i, (images, labels) in enumerate(train_loader):

        images = images.reshape(-1, 28*28)
        
        # 前向传递
        outputs = model(images)
        loss = criterion(outputs, labels)
        
        # 后向传播和优化
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()
        
        if (i+1) % 100 == 0:
            print ('Epoch [{}/{}], Step [{}/{}], Loss: {:.4f}' 
                   .format(epoch+1, num_epochs, i+1, total_step, loss.item()))


# 进行模型测试，测试阶段不再需要计算梯度
with torch.no_grad():
    correct = 0
    total = 0
    for images, labels in test_loader:
        images = images.reshape(-1, 28*28)
        outputs = model(images)
        print(outputs.data.size())
        _, predicted = torch.max(outputs.data, 1)
        total += labels.size(0)
        correct += (predicted == labels).sum()

    print('Accuracy of the model on the 10000 test images: {} %'.format(100 * correct / total))

# 保存模型
torch.save(model.state_dict(), 'model.ckpt')

模型最终准确率：89%     另外，提高模型迭代次数可以进一步提高准确率