Python打卡DAY34

DAY34：GPU训练及类的call方法

恩师@浙大疏锦行

知识点：

一、查看CPU性能

import wmi

c = wmi.WMI()
processors = c.Win32_Processor()

for processor in processors:
    print(f"CPU 型号: {processor.Name}")
    print(f"核心数: {processor.NumberOfCores}")
    print(f"线程数: {processor.NumberOfLogicalProcessors}")

二、用GPU训练模型

1、设置GPU设备

# 设置GPU设备
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
print(f"使用设备: {device}")

2、准备数据

# 加载鸢尾花数据集
iris = load_iris()
X = iris.data  # 特征数据
y = iris.target  # 标签数据

# 划分训练集和测试集
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# 归一化数据
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

# 将数据转换为PyTorch张量并移至GPU
# 分类问题交叉熵损失要求标签为long类型
# 张量具有to(device)方法，可以将张量移动到指定的设备上
X_train = torch.FloatTensor(X_train).to(device)
y_train = torch.LongTensor(y_train).to(device)
X_test = torch.FloatTensor(X_test).to(device)
y_test = torch.LongTensor(y_test).to(device)

3、设计神经网络结构

class MLP(nn.Module):
    def __init__(self):
        super(MLP, self).__init__()
        self.fc1 = nn.Linear(4, 10)
        self.relu = nn.ReLU()
        self.fc2 = nn.Linear(10, 3)

    def forward(self, x):
        out = self.fc1(x)
        out = self.relu(out)
        out = self.fc2(out)
        return out

# 实例化模型并移至GPU
# MLP继承nn.Module类，所以也具有to(device)方法
model = MLP().to(device)

4、定义损失函数和优化器

# 定义损失函数和优化器
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(model.parameters(), lr=0.01)

# 训练模型
num_epochs = 20000
losses = []

5、模型训练

start_time = time.time()

for epoch in range(num_epochs):
    # 前向传播
    outputs = model(X_train)
    loss = criterion(outputs, y_train)

    # 反向传播和优化
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()

    # 记录损失值
    losses.append(loss.item())

    # 打印训练信息
    if (epoch + 1) % 100 == 0:
        print(f'Epoch [{epoch+1}/{num_epochs}], Loss: {loss.item():.4f}')

time_all = time.time() - start_time
print(f'Training time: {time_all:.2f} seconds')

# 可视化损失曲线
plt.plot(range(num_epochs), losses)
plt.xlabel('Epoch')
plt.ylabel('Loss')
plt.title('Training Loss over Epochs')
plt.show()

三、优化思路

1. 直接不打印训练过程的loss了，但是这样会没办法记录最后的可视化图片，只能肉眼观察loss数值变化。

2. 每隔200个epoch保存一下loss，不需要20000个epoch每次都打印。

四、_call_方法

1、不带参数的call方法

class Counter:
    def __init__(self):
        self.count = 0
    
    def __call__(self):
        self.count += 1
        return self.count

# 使用示例
counter = Counter()
print(counter())  # 输出: 1
print(counter())  # 输出: 2
print(counter.count)  # 输出: 2

2、带参数的_call_方法

class Adder:
    def __call__(self, a, b):
        print("加法")
        return a + b

adder = Adder()
print(adder(3, 5))  # 输出: 8