• Pytorch学习之源码理解:pytorch/examples/mnists


    Pytorch学习之源码理解:pytorch/examples/mnists

    from __future__ import print_function
    import argparse
    import torch
    import torch.nn as nn
    import torch.nn.functional as F
    import torch.optim as optim
    from torchvision import datasets, transforms
    from torch.optim.lr_scheduler import StepLR
    
    
    class Net(nn.Module):
        def __init__(self):
            super(Net, self).__init__()#调用父类的构造方法
            self.conv1 = nn.Conv2d(1, 32, 3, 1)#输入1个channel,输出32个channels,kernel_size=3,stride(步长)=1
            self.conv2 = nn.Conv2d(32, 64, 3, 1)#再变成64channels
            self.dropout1 = nn.Dropout2d(0.25)#以0.25的概率dropout
            self.dropout2 = nn.Dropout2d(0.5)
            self.fc1 = nn.Linear(9216, 128)#9216->128
            self.fc2 = nn.Linear(128, 10)
        #定义网络各层
        def forward(self, x):
            x = self.conv1(x)
            #线性整流函数(Rectified Linear Unit, ReLU)是一个激活函数,这是当成一层了
            #卷积神经网络中,若不采用非线性激活,会导致神经网络只能拟合线性可分的数据,因此通常会在卷积操作后,添加非线性激活单元,其中包括logistic-sigmoid、tanh-sigmoid、ReLU等。
            x = F.relu(x)
            x = self.conv2(x)
            x = F.max_pool2d(x, 2)
            x = self.dropout1(x)
            x = torch.flatten(x, 1)
            x = self.fc1(x)
            x = F.relu(x)
            x = self.dropout2(x)
            x = self.fc2(x)
            output = F.log_softmax(x, dim=1)
            return output
    
    
    def train(args, model, device, train_loader, optimizer, epoch):
        model.train()
        #这是两种模式
        #model.train() :启用 BatchNormalization 和 Dropout
        #model.eval() :不启用 BatchNormalization 和 Dropout
        #model.eval(),pytorch会自动把BN和DropOut固定住,不会取平均,而是用训练好的值。
        # 不然的话,一旦test的batch_size过小,很容易就会被BN层导致生成图片颜色失真极大;在模型测试阶段使用
        #trainloader对每一个batch加了id
        for batch_idx, (data, target) in enumerate(train_loader):
            #读入数据到device中,之后就用新的变量表示就可,对程序不影响(物理层和应用层)
            data, target = data.to(device), target.to(device)
            optimizer.zero_grad()#初始化优化器参数
            output = model(data)
            loss = F.nll_loss(output, target)#计算loss
            loss.backward()#反向传播,计算梯度和
            optimizer.step()#调整参数。
            #上面的方法都是共享一个参数空间的,所以不需要传递参数
            if batch_idx % args.log_interval == 0:
                print('Train Epoch: {} [{}/{} ({:.0f}%)]	Loss: {:.6f}'.format(
                    epoch, batch_idx * len(data), len(train_loader.dataset),
                    100. * batch_idx / len(train_loader), loss.item()))
    
    
    def test(args, model, device, test_loader):
        model.eval()
        test_loss = 0
        correct = 0
        with torch.no_grad():
            for data, target in test_loader:
                data, target = data.to(device), target.to(device)
                output = model(data)
                test_loss += F.nll_loss(output, target, reduction='sum').item()  # sum up batch loss
                pred = output.argmax(dim=1, keepdim=True)  # get the index of the max log-probability
                correct += pred.eq(target.view_as(pred)).sum().item()
    
        test_loss /= len(test_loader.dataset)
    
        print('
    Test set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)
    '.format(
            test_loss, correct, len(test_loader.dataset),
            100. * correct / len(test_loader.dataset)))
    
    
    def main():
        # Training settings
        #都是可选参数,是为了调参用的
        parser = argparse.ArgumentParser(description='PyTorch MNIST Example')#加上参数描述,在--help中输出
        parser.add_argument('--batch-size', type=int, default=64, metavar='N',
                            help='input batch size for training (default: 64)')
        parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
                            help='input batch size for testing (default: 1000)')
        parser.add_argument('--epochs', type=int, default=14, metavar='N',
                            help='number of epochs to train (default: 14)')
        parser.add_argument('--lr', type=float, default=1.0, metavar='LR',
                            help='learning rate (default: 1.0)')
        parser.add_argument('--gamma', type=float, default=0.7, metavar='M',
                            help='Learning rate step gamma (default: 0.7)')
        parser.add_argument('--no-cuda', action='store_true', default=False,
                            help='disables CUDA training')
        parser.add_argument('--seed', type=int, default=1, metavar='S',
                            help='random seed (default: 1)')
        parser.add_argument('--log-interval', type=int, default=10, metavar='N',
                            help='how many batches to wait before logging training status')
    
        parser.add_argument('--save-model', action='store_true', default=False,
                            help='For Saving the current Model')
        args = parser.parse_args()#获取参数,从这里就可以开始调用这些参数了。没有输入也没有设置默认值的就是null,用在布尔表达式里面也可以表示false
        use_cuda = not args.no_cuda and torch.cuda.is_available()#有cuda并且没设置参数说不用才用cuda
    
        torch.manual_seed(args.seed)#设置随机种子,以便于生成随机数
    
        device = torch.device("cuda" if use_cuda else "cpu")#决定用cpu还是GPU
    
        kwargs = {'num_workers': 1, 'pin_memory': True} if use_cuda else {}
        #载入训练集
        train_loader = torch.utils.data.DataLoader(
            #torchvision下的datasets模块,如果没发现本地有这个包就下载
            datasets.MNIST('../data', train=True, download=True,
                           transform=transforms.Compose([
                               transforms.ToTensor(),#输出tensor类型
                               transforms.Normalize((0.1307,), (0.3081,))#do normalize
                           ])),
            batch_size=args.batch_size, shuffle=True, **kwargs)#一次读多少
        #载入测试集
        test_loader = torch.utils.data.DataLoader(
            datasets.MNIST('../data', train=False, transform=transforms.Compose([
                               transforms.ToTensor(),
                               transforms.Normalize((0.1307,), (0.3081,))
                           ])),
            batch_size=args.test_batch_size, shuffle=True, **kwargs)
        #将模型读入device
        model = Net().to(device)
        #设置优化器,这里使用的是Adagrad优化方法(Adaptive Gradient)
        optimizer = optim.Adadelta(model.parameters(), lr=args.lr)
        #等间隔调整学习率 StepLR
        #等间隔调整学习率,调整倍数为 gamma 倍,调整间隔为 step_size。间隔单位是step。需要注意的是, step 通常是指 epoch,不要弄成 iteration 了。
        scheduler = StepLR(optimizer, step_size=1, gamma=args.gamma)
        for epoch in range(1, args.epochs + 1):#迭代次数
            train(args, model, device, train_loader, optimizer, epoch)
            test(args, model, device, test_loader)
            scheduler.step()#每次迭代之后调整学习率
    
        if args.save_model:#保存模型
            torch.save(model.state_dict(), "mnist_cnn.pt")
    
    
    if __name__ == '__main__':
        main()
    
  • 相关阅读:
    Zookeeper实现配置中心
    Springboot
    分布式事务框架
    Spring 事件Application Event
    了解“事务机制” 过程
    mysql : 行锁,表锁,共享锁,排他锁,悲观锁,乐观锁
    数据库隔离级别和实现原理
    aop用于日志
    exceptionHandler統一处理异常
    mysql 点总结
  • 原文地址:https://www.cnblogs.com/jiading/p/11946408.html
Copyright © 2020-2023  润新知