多层感知机（MLP）

是什么？

导入需要的包

import torch.nn as nn
import torch
import sys

sys.path.append(".")
import d2lzh_pytorch as d2l
from torch.nn import init

定义模型

不同于 softmax 和简单的线性回归，多层感知机多了隐藏层，由于层数的增多，需要在不同层之间定义激活函数，否则多个线性层最终能合并成一个线性层。

定义隐藏层权重矩阵为784x256，定义输出层权重矩阵为256x10

num_inputs, num_outputs, num_hidden = 784, 10, 256

net = nn.Sequential(
    d2l.FlattenLayer(),
    nn.Linear(num_inputs, num_hidden),
    nn.ReLU(),
    nn.Linear(num_hidden, num_outputs),
)

初始化模型参数

for parameter in net.parameters():
    init.normal_(parameter, 0, 0.01)

读取数据、定义损失函数和优化器并训练模型

这里为了优化代码，直接调用了 d2lzh 包中的train_ch3函数。

batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
loss = nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(net.parameters(), lr=0.1)
num_epochs = 5
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, batch_size, None, None, optimizer)

d2lzh包中的train_ch3函数定义如下：

`train_ch3`函数定义

def train_ch3(net, train_iter, test_iter, loss, num_epochs, batch_size,
              params=None, lr=None, optimizer=None):
    for epoch in range(num_epochs):
        train_l_sum, train_acc_sum, n = 0.0, 0.0, 0
        for X, y in train_iter:
            y_hat = net(X)
            l = loss(y_hat, y).sum()

            # 梯度清零
            if optimizer is not None:
                optimizer.zero_grad()
            elif params is not None and params[0].grad is not None:
                for param in params:
                    param.grad.data.zero_()

            l.backward()
            if optimizer is None:
                sgd(params, lr, batch_size)
            else:
                optimizer.step()


            train_l_sum += l.item()
            train_acc_sum += (y_hat.argmax(dim=1) == y).sum().item()
            n += y.shape[0]
        test_acc = evaluate_accuracy(test_iter, net)
        print('epoch %d, loss %.4f, train acc %.3f, test acc %.3f'
              % (epoch + 1, train_l_sum / n, train_acc_sum / n, test_acc))
ef train_ch3(net, train_iter, test_iter, loss, num_epochs, batch_size,
              params=None, lr=None, optimizer=None):
    for epoch in range(num_epochs):
        train_l_sum, train_acc_sum, n = 0.0, 0.0, 0
        for X, y in train_iter:
            y_hat = net(X)
            l = loss(y_hat, y).sum()

            # 梯度清零
            if optimizer is not None:
                optimizer.zero_grad()
            elif params is not None and params[0].grad is not None:
                for param in params:
                    param.grad.data.zero_()

            l.backward()
            if optimizer is None:
                sgd(params, lr, batch_size)
            else:
                optimizer.step()

            train_l_sum += l.item()
            train_acc_sum += (y_hat.argmax(dim=1) == y).sum().item()
            n += y.shape[0]
        test_acc = evaluate_accuracy(test_iter, net)
        print('epoch %d, loss %.4f, train acc %.3f, test acc %.3f'
              % (epoch + 1, train_l_sum / n, train_acc_sum / n, test_acc))