关于pytorch训练和验证图片的简单程序历程

这个是关于pytorch训练图片和验证图片数据集的简单程序历程

这个程序是学习b站上土堆的pytorch教程而写下的程序，欢迎大家学习土堆的有关教程，讲的挺好的。

import torchvision
from torch import nn
from torch.nn import Conv2d, MaxPool2d, Flatten, Linear
from torch.utils.data import DataLoader
from torch.utils.tensorboard import SummaryWriter

from model import *

train_data = torchvision.datasets.CIFAR10(root='dataset', train=True, transform=torchvision.transforms.ToTensor(),
                                          download=True)

test_data = torchvision.datasets.CIFAR10(root='dataset', train=False, transform=torchvision.transforms.ToTensor(),
                                         download=True)

train_data_dataloader = DataLoader(train_data, batch_size=64)
test_data_dataloader = DataLoader(test_data, batch_size=64)

total_test = len(test_data)

class Tudui(nn.Module):
    def __init__(self):
        super(Tudui, self).__init__()
        self.s1 = nn.Sequential(
            Conv2d(3,32,5,1,2),
            MaxPool2d(2),
            Conv2d(32, 32, 5,1, 2),
            MaxPool2d(2),
            Conv2d(32, 64, 5,1, 2),
            MaxPool2d(2),
            Flatten(),
            Linear(1024, 64),
            Linear(64, 10)
        )

    def forward(self, x):
        x = self.s1(x)
        return x

total_train_step = 0
total_test_step = 0
tudui = Tudui()

writer = SummaryWriter('logs_fn')
step = 0

loss_fn = nn.CrossEntropyLoss()
learning_rate = 1e-2
optimizer = torch.optim.SGD(tudui.parameters(), lr=learning_rate)
epochs = 20

for i in range(epochs):

    tudui.train()
    for data in train_data_dataloader:
        imgs, tars = data
        output = tudui(imgs)
        loss = loss_fn(output, tars)
        optimizer.zero_grad()

        loss.backward()
        optimizer.step()

        total_train_step = total_train_step + 1

        if total_train_step % 100 == 0:
            print("第{}轮训练,损失值是：{}".format(total_train_step, loss.item()))
            writer.add_scalar('train', loss, total_train_step)

    # 测试开始
    tudui.eval()
    total_test_loss = 0
    total_accuracy = 0
    with torch.no_grad():
        for data in test_data_dataloader:
            imgs, tars = data
            output = tudui(imgs)

            loss = loss_fn(output, tars)
            total_test_loss = total_test_loss + loss.item()
            accuracy = (output.argmax(1) == tars).sum()
            total_accuracy = total_accuracy + accuracy

    print('第{}测试，损失值是：{}'.format(total_test_step, total_test_loss))
    writer.add_scalar('test', total_test_loss, step)
    print('正确率是{}'.format(total_accuracy / total_test))
    writer.add_scalar('accuracy', total_accuracy, total_test_step)
    step = step + 1
    total_test_step = total_test_step + 1

    torch.save(tudui, "tudui_{}.pth".format(i))
    print('模型已保存')

writer.close()

这个是一个简单的程序