Inception 实现MNIST识别

神经网络架构如下：

实现代码如下，运算速度有点长。

import torch
from torchvision import datasets, transforms
from torch.utils.data import DataLoader
import torch.nn.functional as F
import matplotlib.pyplot as plt

batch_size = 64
learning_rate = 0.01

#将图像转化为张量
transform = transforms.Compose([ #把读取的图片进行张量化
    transforms.ToTensor(),
    transforms.Normalize((0.1307,), (0.3081,))
])

#加载数据集
train_dataset = datasets.MNIST(root='../dataset',
                               train=True,
                               transform=transform,
                               download=False)

train_loader = DataLoader(train_dataset,
                          batch_size=batch_size,
                          shuffle=True)

test_dataset = datasets.MNIST(root='../dataset',
                              download=False,
                              train=True,
                              transform=transform)

test_loader = DataLoader(test_dataset,
                         shuffle=False,
                         batch_size=batch_size)

#定义带有卷积层的神经网络
class InceptionA(torch.nn.Module):
    def __init__(self,in_channels):
        super(InceptionA, self).__init__()
        self.branch1x1 = torch.nn.Conv2d(in_channels=in_channels,out_channels=16,kernel_size=1,bias=False)

        self.branch_pooling = torch.nn.Conv2d(in_channels=in_channels,out_channels=24,kernel_size=1,bias=False)

        self.branch5x5_1 = torch.nn.Conv2d(in_channels=in_channels,out_channels=16,kernel_size=1,bias=False)
        self.branch5x5_2 = torch.nn.Conv2d(in_channels=16,out_channels=24,kernel_size=5,bias=False,padding=2)

        self.branch3x3_1 = torch.nn.Conv2d(in_channels=in_channels,out_channels=16,kernel_size=1,bias=False)
        self.branch3x3_2 = torch.nn.Conv2d(in_channels=16,out_channels=24,kernel_size=3,bias=False,padding=1)
        self.branch3x3_3 = torch.nn.Conv2d(in_channels=24,out_channels=24,kernel_size=3,bias=False,padding=1)

    def forward(self,x):
        branch1x1 = self.branch1x1(x)

        branch_pooling = F.avg_pool2d(x,kernel_size=3,stride=1,padding=1)
        branch_pooling = self.branch_pooling(branch_pooling)

        branch5x5 = self.branch5x5_1(x)
        branch5x5 = self.branch5x5_2(branch5x5)

        branch3x3 = self.branch3x3_1(x)
        branch3x3 = self.branch3x3_2(branch3x3)
        branch3x3 = self.branch3x3_3(branch3x3)

        output = [branch1x1, branch_pooling, branch5x5, branch3x3]

        return torch.cat(output, 1)

class Net(torch.nn.Module):
    def __init__(self):
        super(Net,self).__init__()
        self.conv1 = torch.nn.Conv2d(in_channels=1,out_channels=10,kernel_size=5,bias=False)
        self.conv2 = torch.nn.Conv2d(in_channels=88,out_channels=20,kernel_size=5,bias=False)

        self.incep1 = InceptionA(in_channels = 10)
        self.incep2 = InceptionA(in_channels = 20)

        self.mp = torch.nn.MaxPool2d(kernel_size=2)
        self.fc = torch.nn.Linear(1408,10)

    def forward(self,x):
        in_size = x.size(0)
        x = F.relu(self.mp(self.conv1(x)))
        x = self.incep1(x)
        x = F.relu(self.mp(self.conv2(x)))
        x = self.incep2(x)
        x = x.view(in_size, -1)
        x = self.fc(x)
        return x

#实例化对象
model = Net()
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
model.to(device)

#定义损失函数和优化器
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(),lr=learning_rate,momentum=0.5)

#训练过程
def train(epoch):
    model.train()
    running_loss = 0.0
    for batch_idx, data in enumerate(train_loader, 0):
        inputs, target = data
        inputs, target = inputs.to(device), target.to(device)
        optimizer.zero_grad()
        outputs = model(inputs)
        loss = criterion(outputs, target)
        loss.backward()
        optimizer.step()
        running_loss += loss.item()

        if batch_idx % 300 == 299:
            print(f'[{epoch + 1}, {batch_idx + 1:5d}] loss: {running_loss / 300:.3f}')
            running_loss = 0.0

#测试过程
def test(epoch):
    model.eval()
    correct = 0
    total = 0
    with torch.no_grad():
        for data in test_loader:
            inputs, labels = data
            inputs, labels = inputs.to(device), labels.to(device)
            outputs = model(inputs)
            _, predicted = torch.max(outputs.data, 1)
            total += labels.size(0)
            correct += (predicted == labels).sum().item()
        accuracy = 100 * correct / total
        print(f'Accuracy on test set: {accuracy:.3f}')
        return accuracy

if __name__ == '__main__':
    epochs = 10
    epoch_list = []
    accuracy_list = []

    for epoch in range(epochs):
        train(epoch)
        accuracy = test(epoch)
        epoch_list.append(epoch + 1)
        accuracy_list.append(accuracy)

    # 绘制准确率曲线
    plt.figure(figsize=(12, 5))
    plt.plot(epoch_list, accuracy_list, 'b-')
    plt.title('Test Accuracy vs. Epochs')
    plt.xlabel('Epoch')
    plt.ylabel('Accuracy (%)')
    plt.grid(True)
    plt.show()

实现效果如下，准确率可以突破到99以上：