李沐-权重衰退

控制模型的容量
1.参数的个数
2.每个参数值得选择范围
一、使用均方范数作为硬性限制

这里得subject to意味着限制得意思。

解释：限制w得取值范围，让模型不要太复杂，去拟合得像蓝色线一样。限制W范围，拟合得像绿色线，模型就不复杂了

二、使用均方范数来柔性限制
惩罚项（penalty）


具体得


参数更新法则

在参数更新得时候，w会衰退得。
总结

又引入了一个超参数λ

代码

%matplotlib inline
import torch
from torch import nn
from d2l import torch as d2l

n_train, n_test, num_inputs, batch_size = 20, 100, 200, 5
true_w, true_b = torch.ones((num_inputs, 1)) * 0.01, 0.05
train_data = d2l.synthetic_data(true_w, true_b, n_train)
train_iter = d2l.load_array(train_data, batch_size)
test_data = d2l.synthetic_data(true_w, true_b, n_test)
test_iter = d2l.load_array(test_data, batch_size, is_train=False)

这里，模型复杂，数据简单，训练集很小，容易过拟合

初始化权重

def init_params():
    w = torch.normal(0, 1, size=(num_inputs, 1), requires_grad=True)
    b = torch.zeros(1, requires_grad=True)
    return [w, b]

定义惩罚项

def l2_penalty(w):
    return torch.sum(w.pow(2)) / 2

训练

def train(lambd):
    w, b = init_params()
    net, loss = lambda X: d2l.linreg(X, w, b), d2l.squared_loss
    num_epochs, lr = 100, 0.003
    animator = d2l.Animator(xlabel='epochs', ylabel='loss', yscale='log',
                            xlim=[5, num_epochs], legend=['train', 'test'])
    for epoch in range(num_epochs):
        for X, y in train_iter:
            with torch.enable_grad():
                # 增加了L2范数惩罚项，广播机制使l2_penalty(w)成为一个长度为`batch_size`的向量。
                l = loss(net(X), y) + lambd * l2_penalty(w)
            l.sum().backward()
            d2l.sgd([w, b], lr, batch_size)
        if (epoch + 1) % 5 == 0:
            animator.add(epoch + 1, (d2l.evaluate_loss(net, train_iter, loss),
                                     d2l.evaluate_loss(net, test_iter, loss)))
    print('w的L2范数是：', torch.norm(w).item())

简洁实现

weight_decay就是λ。权重衰退的超参数

def train_concise(wd):
    net = nn.Sequential(nn.Linear(num_inputs, 1))
    for param in net.parameters():
        param.data.normal_()
    loss = nn.MSELoss()
    num_epochs, lr = 100, 0.003
    # 偏置参数没有衰减。
    trainer = torch.optim.SGD([{
        "params": net[0].weight,
        'weight_decay': wd}, {
            "params": net[0].bias}], lr=lr)
    animator = d2l.Animator(xlabel='epochs', ylabel='loss', yscale='log',
                            xlim=[5, num_epochs], legend=['train', 'test'])
    for epoch in range(num_epochs):
        for X, y in train_iter:
            with torch.enable_grad():
                trainer.zero_grad()
                l = loss(net(X), y)
            l.backward()
            trainer.step()
        if (epoch + 1) % 5 == 0:
            animator.add(epoch + 1, (d2l.evaluate_loss(net, train_iter, loss),
                                     d2l.evaluate_loss(net, test_iter, loss)))
    print('w的L2范数：', net[0].weight.norm().item())