Pytorch中怎么使用vis.line 来画train_loss train_acc test_acc 这三条线在同一个图中 并给出一个在深度学习中以网络为resnet50 、数据集为cifar10、把输出的模型保存、输出的数据保存到csv文件中、把代码模块化的实战代码

首先，我们需要安装visdom库，可以使用以下命令：

!pip install visdom

接下来，需要在代码中引入相关库和模块：

import torch
import torch.nn as nn
import torch.optim as optim
import torchvision
import torchvision.transforms as transforms
import visdom
import csv

然后，定义一个函数来加载数据集：

def load_dataset():
    transform_train = transforms.Compose(
        [transforms.RandomCrop(32, padding=4),
         transforms.RandomHorizontalFlip(),
         transforms.ToTensor(),
         transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])

    transform_test = 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_train)
    trainloader = torch.utils.data.DataLoader(trainset, batch_size=128,
                                              shuffle=True, num_workers=2)

    testset = torchvision.datasets.CIFAR10(root='./data', train=False,
                                           download=True, transform=transform_test)
    testloader = torch.utils.data.DataLoader(testset, batch_size=128,
                                             shuffle=False, num_workers=2)

    classes = ('plane', 'car', 'bird', 'cat',
               'deer', 'dog', 'frog', 'horse', 'ship', 'truck')

    return trainloader, testloader

接下来，定义一个函数来构建ResNet50网络：

class ResNet50(nn.Module):
    def __init__(self, num_classes=10):
        super(ResNet50, self).__init__()
        self.resnet50 = torchvision.models.resnet50(pretrained=False)
        num_ftrs = self.resnet50.fc.in_features
        self.resnet50.fc = nn.Linear(num_ftrs, num_classes)

    def forward(self, x):
        x = self.resnet50(x)
        return x

然后，定义一个函数来训练模型：

def train(model, trainloader, criterion, optimizer, epoch, device, vis):
    model.train()
    train_loss = 0.0
    correct = 0
    total = 0
    for batch_idx, (inputs, targets) in enumerate(trainloader):
        inputs, targets = inputs.to(device), targets.to(device)
        optimizer.zero_grad()
        outputs = model(inputs)
        loss = criterion(outputs, targets)
        loss.backward()
        optimizer.step()

        train_loss += loss.item()
        _, predicted = outputs.max(1)
        total += targets.size(0)
        correct += predicted.eq(targets).sum().item()

        if batch_idx % 10 == 0:
            vis.line(X=torch.Tensor([epoch+batch_idx/len(trainloader)]), Y=torch.Tensor([train_loss/len(trainloader)]), win='train_loss', update='append', name='train_loss')
            vis.line(X=torch.Tensor([epoch+batch_idx/len(trainloader)]), Y=torch.Tensor([100.*correct/total]), win='train_acc', update='append', name='train_acc')

    print('Train Epoch: {} Loss: {:.6f} Acc: {:.6f}'.format(
        epoch, train_loss/len(trainloader), 100.*correct/total))

接下来，定义一个函数来测试模型：

def test(model, testloader, criterion, epoch, device, vis):
    model.eval()
    test_loss = 0.0
    correct = 0
    total = 0
    with torch.no_grad():
        for batch_idx, (inputs, targets) in enumerate(testloader):
            inputs, targets = inputs.to(device), targets.to(device)
            outputs = model(inputs)
            loss = criterion(outputs, targets)

            test_loss += loss.item()
            _, predicted = outputs.max(1)
            total += targets.size(0)
            correct += predicted.eq(targets).sum().item()

            if batch_idx % 10 == 0:
                vis.line(X=torch.Tensor([epoch+batch_idx/len(testloader)]), Y=torch.Tensor([100.*correct/total]), win='test_acc', update='append', name='test_acc')

    print('Test Epoch: {} Loss: {:.6f} Acc: {:.6f}'.format(
        epoch, test_loss/len(testloader), 100.*correct/total))

接下来，定义主函数来调用以上函数：

def main():
    vis = visdom.Visdom()
    device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
    trainloader, testloader = load_dataset()
    model = ResNet50().to(device)
    criterion = nn.CrossEntropyLoss()
    optimizer = optim.SGD(model.parameters(), lr=0.1, momentum=0.9, weight_decay=5e-4)

    for epoch in range(100):
        train(model, trainloader, criterion, optimizer, epoch, device, vis)
        test(model, testloader, criterion, epoch, device, vis)

    torch.save(model.state_dict(), 'resnet50_cifar10.pth')

    with open('output.csv', mode='w') as file:
        writer = csv.writer(file)
        writer.writerow(['epoch', 'train_loss', 'train_acc', 'test_acc'])
        for epoch in range(100):
            train_loss = vis.line(Y=torch.Tensor([0]), win='train_loss', name='train_loss', update='none')
            train_acc = vis.line(Y=torch.Tensor([0]), win='train_acc', name='train_acc', update='none')
            test_acc = vis.line(Y=torch.Tensor([0]), win='test_acc', name='test_acc', update='none')
            writer.writerow([epoch+1, train_loss['Y'][-1], train_acc['Y'][-1], test_acc['Y'][-1]])

最后，运行主函数即可进行训练、测试、模型保存和输出数据到csv文件的操作：

if __name__ == '__main__':
    main()

完整代码如下：