pytorch-------Cnn_mnist

pytorch1.1.0 + torchvision0.3.0 + cuda10.0.130 + NVIDIA-SMI 470.103.01
import torch
from torchvision import transforms #将图像转化为张量
from torchvision import datasets  #对数据集相关处理
from torch.utils.data import DataLoader #下载数据集
import torch.nn.functional as F  #使用relu()函数
import torch.optim as optim #优化器
batch_size = 64
"将图像转化成张量:28X28--->1X28X28 CXHXW,并归一化到0-1,0-255--->0-1;0.1307:均值,0.3081:标准差"
transform = transforms.Compose([transforms.ToTensor(),transforms.Normalize((0.1307,),(0.3081,))])
train_dataset = datasets.MNIST(root='../dataset/mnist',train=True,transform=transform,download=True)
test_dataset = datasets.MNIST(root='../dataset/mnist',train=False,transform=transform,download=True)
train_loader = DataLoader(dataset=train_dataset,batch_size=batch_size,shuffle=True)
test_loader = DataLoader(dataset=test_dataset,batch_size=batch_size,shuffle=False)
"2.模型设计"
"输入 NX1X28X28变为 NX784"
"输出为NX10"
class Net(torch.nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = torch.nn.Conv2d(1,10,kernel_size=5)
        self.conv2 = torch.nn.Conv2d(10,20,kernel_size=5)
        self.pooling = torch.nn.MaxPool2d(2)
        self.fc = torch.nn.Linear(320,10)
    def forward(self,x):
       batch_size = x.size(0)
       x = F.relu(self.pooling(self.conv1(x)))
       #x = self.pooling(F.relu(self.conv1(x)))
       x = F.relu(self.pooling(self.conv2(x)))
       #拉长数据:(n,1,28,28)---->(n,784)
       x = x.view(batch_size,-1)
       x = self.fc(x)
       return x
model = Net()
#该句表示使用GPU,若无GPU,可直接删除
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
model.to(device)
"3.优化器和损失"
criterion = torch.nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(),lr=0.01,momentum=0.5)
"4.训练和测试"
def train(epoch):
    running_loss = 0.0
    for batch_idx,data in enumerate(train_loader,0):
        inputs,target = data
        #该句表示使用GPU,若无GPU,可直接删除
        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:  #每300轮输出一次
            print('[%d,%5d] loss: %.3f'%(epoch+1, batch_idx+1,running_loss/300))
            running_loss = 0.0
def test():
    correct = 0
    total = 0
    with torch.no_grad():
        for data  in test_loader:
            images,labels = data
            #该句表示使用GPU,若无GPU,可直接删除
            images, labels = images.to(device), labels.to(device)
            outputs = model(images)
            # 取出每一行最大值的下标(0-9),dim=1:第一个维度(行);dim=0:第0个维度,列
            _,predicted = torch.max(outputs .data,dim=1)

            total += labels.size(0) #取NX1矩阵中的N  ==batch_size
            correct += (predicted == labels).sum().item() #是否为真
    #print(correct,total)
    print("Accuracy on test set: %d %%"%(100*correct/total)) #正确的/总数

if __name__ == '__main__':
    for epoch in range(10):
        train(epoch)
        test()
上一篇:图像分类数据集的读取


下一篇:大概是全网最详细的何恺明团队顶作MoCo系列解读...(完结篇)