在多分类任务实验中 用torch.nn实现 𝑳𝟐 正则化

1 导入实验所需要的包

import torch
import torch.nn as nn
import numpy as np
import torchvision
import torchvision.transforms as transforms
import matplotlib.pyplot as plt
%matplotlib inline

2 下载MNIST数据集以及读取数据

train_dataset = torchvision.datasets.MNIST(root='../Datasets/MNIST', train=True, transform=transforms.ToTensor(),download=True)
test_dataset = torchvision.datasets.MNIST(root='../Datasets/MNIST', train=False, transform = transforms.ToTensor(),download=True)
train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=32, shuffle=True)
test_loader = torch.utils.data.DataLoader(test_dataset, batch_size=32, shuffle=False)

3 初始化参数

num_inputs, num_outputs, num_hiddens = 784, 10, 600
def init_w_b():
    W1 = torch.tensor(np.random.normal(0, 0.01, (num_inputs, num_hiddens)), dtype=torch.float)
    b1 = torch.zeros(num_hiddens, dtype = torch.float)
    W2 = torch.tensor(np.random.normal(0, 0.01, (num_hiddens, num_outputs)), dtype=torch.float)
    b2 = torch.zeros(num_outputs,dtype=torch.float)

    params = [W1,b1,W2,b2]
    for param in params:
        param.requires_grad_(requires_grad=True)
    return W1,b1,W2,b2

4 定义模型

class LinearNet(nn.Module):
    def __init__(self,num_inputs, num_outputs, num_hiddens):
        super(LinearNet,self).__init__()
        self.linear1 = nn.Linear(num_inputs,num_hiddens)
        self.relu = nn.ReLU()
        self.linear2 = nn.Linear(num_hiddens,num_outputs)
        self.flatten  = nn.Flatten()
    
    def forward(self,x):
        x = self.flatten(x)
        x = self.linear1(x)
        x = self.relu(x)
        x = self.linear2(x)
        y = self.relu(x)
        return y

5 定义训练函数

def train_torch(lamda):
    num_epochs = 20
    train_ls, test_ls = [], []
    for epoch in range(num_epochs):
        ls, count = 0, 0
        for X,y in train_loader:
            l=loss(net(X),y)
            optimizer_w.zero_grad()
            optimizer_b.zero_grad()
            l.backward()
            optimizer_w.step()
            optimizer_b.step()
            ls += l.item()
            count += y.shape[0]
        train_ls.append(ls)
        ls, count = 0, 0
        for X,y in test_loader:
            l=loss(net(X),y)
            ls += l.item()
            count += y.shape[0]
        test_ls.append(ls)
        
        print('epoch: %d, train loss: %f, test loss: %f'%(epoch+1,train_ls[-1],test_ls[-1]))
    return train_ls,test_ls

6 开始训练

Lamda = [0,0.1,0.2,0.3,0.4,0.5]
torch_Train_ls, torch_Test_ls = [], []
for lamda in Lamda:
    W1,b1,W2,b2 = init_w_b()
    loss = nn.CrossEntropyLoss()
    net = LinearNet(num_inputs, num_outputs, num_hiddens)
    optimizer_w = torch.optim.SGD([W1,W2],lr = 0.001,weight_decay=lamda)
    optimizer_b = torch.optim.SGD([b1,b2],lr = 0.001)
    train_ls, test_ls = train_torch(lamda)
    torch_Train_ls.append(train_ls)
    torch_Test_ls.append(test_ls)

7 绘制训练集和测试集的loss曲线

x = np.linspace(0,len(torch_Train_ls[1]),len(torch_Train_ls[1]))
plt.figure(figsize=(10,8))
for i in range(0,len(Lamda)):
    plt.plot(x,torch_Train_ls[i],label= f'L2_Regularization:{Lamda [i]}',linewidth=1.5)
    plt.xlabel('epoch')
    plt.ylabel('loss')
plt.legend(loc=2, bbox_to_anchor=(1.1,1.0),borderaxespad = 0.)
plt.title('train loss')
plt.show()