Pytorch中使用显卡（GPU）训练模型

概述

在进行深度学习训练模型时，对于计算量小一些的模型，是可以在CPU上进行的。但是当计算量比较大时，我们希望利用GPU并行计算的能力去加快训练的速度。

方法步骤

Pytroch中使用GPU训练模型需要以下四步：

1. 创建模型
2. 定义device
3. 将模型加载到GPU
4. 将输入和输出加载到GPU

---

对应的代码为：

# 1.创建模型
model = Net()
# 2.定义device
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
# 3.将模型加载到GPU(所定义的device) 
model.to(device)
# 4.将输入和输出加载到GPU
inputs, target = inputs.to(device), target.to(device)

---

示例说明

在下面这个示例中给出每一步如何操作：
原来用CPU的版本：

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):
            # Flatten data from (n, 1, 28, 28) to (n, 784)
            batch_size = x.size(0)
            x = F.relu(self.pooling(self.conv1(x)))
            x = F.relu(self.pooling(self.conv2(x)))
            x = x.view(batch_size, -1) # flatten
            x = self.fc(x)
            return x
        
model = Net()

def train(epoch):
    running_loss = 0.0
    for batch_idx, data in enumerate(train_loader, 0):
        inputs, target = data
        optimizer.zero_grad()
        # forward + backward + update
        outputs = model(inputs)
        loss = criterion(outputs, target)
        loss.backward()
        optimizer.step()
        running_loss += loss.item()
        if batch_idx % 300 == 299:
            print('[%d, %5d] loss: %.3f' % (epoch + 1, batch_idx + 1, running_loss / 2000))
            running_loss = 0.0

def test():
    correct = 0
    total = 0
    with torch.no_grad():
        for data in test_loader:
            inputs, target = data
            outputs = model(inputs)
            _, predicted = torch.max(outputs.data, dim=1)
            total += target.size(0)
            correct += (predicted == target).sum().item()
            print('Accuracy on test set: %d %% [%d/%d]' % (100 * correct / total, correct, total))

---

现在我们使用后上面给出的四步将模型加载到GPU：

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):
            # Flatten data from (n, 1, 28, 28) to (n, 784)
            batch_size = x.size(0)
            x = F.relu(self.pooling(self.conv1(x)))
            x = F.relu(self.pooling(self.conv2(x)))
            x = x.view(batch_size, -1) # flatten
            x = self.fc(x)
            return x
        
model = Net() # 1.创建模型，这个CPU版本的也需要

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu") # 2. 定义device（GPU）
model.to(device) #将模型加载到GPU

def train(epoch):
    running_loss = 0.0
    for batch_idx, data in enumerate(train_loader, 0):
        # inputs, target = data
        inputs, target = inputs.to(device), target.to(device) # 4. 将输入和输出加载到GPU
        optimizer.zero_grad()
        # forward + backward + update
        outputs = model(inputs)
        loss = criterion(outputs, target)
        loss.backward()
        optimizer.step()
        running_loss += loss.item()
        if batch_idx % 300 == 299:
            print('[%d, %5d] loss: %.3f' % (epoch + 1, batch_idx + 1, running_loss / 2000))
            running_loss = 0.0

def test():
    correct = 0
    total = 0
    with torch.no_grad():
        for data in test_loader:
            # inputs, target = data
            inputs, target = inputs.to(device), target.to(device) # 4. 将输入和输出加载到GPU
            outputs = model(inputs)
            _, predicted = torch.max(outputs.data, dim=1)
            total += target.size(0)
            correct += (predicted == target).sum().item()
            print('Accuracy on test set: %d %% [%d/%d]' % (100 * correct / total, correct, total))

---

参考资料

[1] https://www.bilibili.com/video/BV1Y7411d7Ys?p=10