PyTorch MNIST
长浔 人气:0前言:
本篇文章基于卷积神经网络CNN,使用PyTorch实现MNIST数据集手写数字识别。
一、PyTorch是什么?
PyTorch 是一个 Torch7 团队开源的 Python 优先的深度学习框架,提供两个高级功能:
- 强大的 GPU 加速 Tensor 计算(类似 numpy)
- 构建基于 tape 的自动升级系统上的深度神经网络
你可以重用你喜欢的 python 包,如 numpy、scipy 和 Cython ,在需要时扩展 PyTorch。
二、程序示例
下面案例可供运行参考
1.引入必要库
import torchvision import torch from torch.utils.data import DataLoader import torch.nn.functional as F
2.下载数据集
这里设置download=True,将会自动下载数据集,并存储在./data文件夹。
train_data = torchvision.datasets.MNIST(root="./data",train=True,transform=torchvision.transforms.ToTensor(),download=True) test_data = torchvision.datasets.MNIST(root="./data",train=False,transform=torchvision.transforms.ToTensor(),download=True)
3.加载数据集
batch_size=32表示每一个batch中包含32张手写数字图片,shuffle=True表示打乱测试集(data和target仍一一对应)
train_loader = DataLoader(train_data,batch_size=32,shuffle=True) test_loader = DataLoader(test_data,batch_size=32,shuffle=False)
4.搭建CNN模型并实例化
class Net(torch.nn.Module): def __init__(self): super(Net,self).__init__() self.con1 = torch.nn.Conv2d(1,10,kernel_size=5) self.con2 = 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.con1(x))) x = F.relu(self.pooling(self.con2(x))) x = x.view(batch_size,-1) x = self.fc(x) return x #模型实例化 model = Net()
5.交叉熵损失函数损失函数及SGD算法优化器
lossfun = torch.nn.CrossEntropyLoss() opt = torch.optim.SGD(model.parameters(),lr=0.01,momentum=0.5)
6.训练函数
def train(epoch): running_loss = 0.0 for i,(inputs,targets) in enumerate(train_loader,0): # inputs,targets = inputs.to(device),targets.to(device) opt.zero_grad() outputs = model(inputs) loss = lossfun(outputs,targets) loss.backward() opt.step() running_loss += loss.item() if i % 300 == 299: print('[%d,%d] loss:%.3f' % (epoch+1,i+1,running_loss/300)) running_loss = 0.0
7.测试函数
def test(): total = 0 correct = 0 with torch.no_grad(): for (inputs,targets) in test_loader: # inputs, targets = inputs.to(device), targets.to(device) outputs = model(inputs) _,predicted = torch.max(outputs.data,dim=1) total += targets.size(0) correct += (predicted == targets).sum().item() print(100*correct/total)
8.运行
if __name__ == '__main__': for epoch in range(20): train(epoch) test()
三、总结
加载全部内容