深度学习教程之基于代码一步一步教你机器学习中多层感知器(MLP)的原理

import torch
import torch.nn as nn
import torch.optim as optim
from torchvision import datasets, transforms
 
# 设置随机种子，以便结果可复现
torch.manual_seed(42)
 
# 定义数据预处理的管道
transform = transforms.Compose([
    transforms.ToTensor(),                           # 将图像转换为张量
    transforms.Normalize((0.5,), (0.5,))             # 归一化
])
 
# 加载并预处理MNIST数据集
trainset = datasets.MNIST(root='./data', train=True, download=True, transform=transform)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=32, shuffle=True)
 
testset = datasets.MNIST(root='./data', train=False, download=True, transform=transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=32, shuffle=False)

# 定义MLP模型
class MLP(nn.Module):
    def __init__(self):
        super(MLP, self).__init__()
        self.fc1 = nn.Linear(28 * 28, 128)             # 全连接层，输入大小为28*28，输出大小为128
        self.relu = nn.ReLU()                          # ReLU激活函数
        self.fc2 = nn.Linear(128, 10)                   # 全连接层，输入大小为128，输出大小为10（类别数）
 
    def forward(self, x):
        x = x.view(x.size(0), -1)                       # 展平输入张量
        x = self.relu(self.fc1(x))
        x = self.fc2(x)
        return x

# 实例化模型和损失函数
net = MLP()
criterion = nn.CrossEntropyLoss()                    # 交叉熵损失函数
optimizer = optim.SGD(net.parameters(), lr=0.01)      # 随机梯度下降优化器

# 训练模型
for epoch in range(5):                               # 进行5个epoch的训练
    running_loss = 0.0
    for i, data in enumerate(trainloader, 0):
        inputs, labels = data
        optimizer.zero_grad()
 
        # 前向传播、反向传播、优化
        outputs = net(inputs)
        loss = criterion(outputs, labels)
        loss.backward()
        optimizer.step()
 
        running_loss += loss.item()
        if i % 1000 == 999:
            print(f'[{epoch + 1}, {i + 1}] loss: {running_loss / 1000:.3f}')
            running_loss = 0.0
 
print('Finished training')

# 在测试集上评估模型
correct = 0
total = 0
with torch.no_grad():
    for data in testloader:
        images, labels = data
        outputs = net(images)
        _, predicted = torch.max(outputs.data, 1)      # 获取每个样本预测的类别索引
        total += labels.size(0)
        correct += (predicted == labels).sum().item()
 
print(f'Accuracy on the test set: {100 * correct / total:.2f}%')