《PyTorch深度学习实践》第九课多分类问题（手写数字MNIST）

b站刘二视频，地址：

《PyTorch深度学习实践》完结合集_哔哩哔哩_bilibili

代码（课中作业）MNIST手写数字识别，采用全连接的方式

运行结果

loss

准确率

代码

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

BATCH_SIZE = 64
TRANSFORM = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))])
                                # tenser类型            归一化
train_set = datasets.MNIST(root='mnist', train=True, transform=TRANSFORM, download=True)
                        #   保存地址       训练集                            若不存在则下载
train_loader = DataLoader(dataset=train_set, shuffle=True, batch_size=BATCH_SIZE)#loader
                          #   设置数据源               乱序
test_set = datasets.MNIST(root='mnist', train=False, transform=TRANSFORM, download=True)
test_loader = DataLoader(dataset=test_set, shuffle=False, batch_size=BATCH_SIZE)

class Net(torch.nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.linear1 = torch.nn.Linear(784, 512)
        self.linear2 = torch.nn.Linear(512, 256)
        self.linear3 = torch.nn.Linear(256, 128)
        self.linear4 = torch.nn.Linear(128, 64)
        self.linear5 = torch.nn.Linear(64, 10)

    def forward(self, x):
        x = x.view(-1, 784)#将其变成向量
        x = F.relu(self.linear1(x))
        x = F.relu(self.linear2(x))
        x = F.relu(self.linear3(x))
        x = F.relu(self.linear4(x))#采用relu做激活函数
        return self.linear5(x) #最后一层不需要做激活

model = Net()

criteration = torch.nn.CrossEntropyLoss()      #交叉熵损失函数
optimizor = torch.optim.SGD(model.parameters(), lr=0.01, momentum=0.5)
                                                        #动量 优化
def train(): #训练
    sum = 0
    for i, data in enumerate(train_loader, 0):
        inputs, lables = data
        y_pred = model(inputs)
        loss = criteration(y_pred, lables)
        sum += loss.item()

        optimizor.zero_grad()
        loss.backward()
        optimizor.step()

        if i % 300 == 299:#每300次输出一次
            sum /= 300
            loss_lst.append(sum)
            sum = 0

def test():
    total = 0
    correct = 0
    with torch.no_grad():
        for i, data in enumerate(test_loader, 0):
            inputs, lables = data
            outputs = model(inputs)
            _, predicted = torch.max(outputs.data, dim=1) #找出每行最大的那个数字（就是预测结果）
            correct += (predicted == lables).sum().item()#正确预测的数量
            total += lables.size(0)
        accurate_lst.append(correct / total * 100)

if __name__ == '__main__':
    loss_lst = []
    accurate_lst = []

    for epoch in range(10):
        train()
        test()

    #可视化画图
    num_lst = [i for i in range(len(loss_lst))]
    plt.plot(num_lst, loss_lst)
    plt.xlabel('i')
    plt.ylabel('loss')
    plt.show()

    num_lst = [i for i in range(len(accurate_lst))]
    plt.plot(num_lst, accurate_lst)
    plt.xlabel('i')
    plt.ylabel('accurate')
    plt.show()