3.10 多层感知机的简洁实现

学习网址：http://tangshusen.me/Dive-into-DL-PyTorch/#/

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

1 定义模型

加了一个全连接层作为隐藏层。它的隐藏单元个数为256，并使用ReLU函数作为激活函数。

num_inputs, num_outputs, num_hiddens = 784, 10, 256

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

for params in net.parameters():
    init.normal_(params, mean=0, std=0.01)

2 读取数据并训练模型

使用的是PyTorch的SGD而不是d2lzh_pytorch里面的sgd，所以就不存在3.9节那样学习率看起来很大的问题了。

batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size)
loss = torch.nn.CrossEntropyLoss()

optimizer = torch.optim.SGD(net.parameters(), lr=0.5)

num_epochs = 5
d2l.train_ch3(net, train_iter, test_iter, loss, num_epochs, batch_size, None, None, optimizer)

epoch 1, loss 0.0031, train acc 0.699, test acc 0.748
epoch 2, loss 0.0019, train acc 0.822, test acc 0.772
epoch 3, loss 0.0017, train acc 0.842, test acc 0.834
epoch 4, loss 0.0015, train acc 0.857, test acc 0.851
epoch 5, loss 0.0014, train acc 0.865, test acc 0.852