pytorch线性回归（构建模型流程）

先贴个笔记：
均方损失函数，注意

SGD更新参数

注意

1.构建模型

设计网络的结构，pytorch有方便的模块和函数可供使用，例如nn.Module，nn.Sequential，nn.Linear：

# 使用类构建
class LinearNet(nn.Module):
    def __init__(self, n_feature):
        super(LinearNet, self).__init__()      # call father function to init 
        self.linear = nn.Linear(n_feature, 1)  # function prototype: `torch.nn.Linear(in_features, out_features, bias=True)`
        '''
        # method 1
        self.linear = nn.Sequential(
                        nn.Linear(n_feature, 1)
                        # nn.Linear(...)
                )

        # method 2
        self.linear = nn.Sequential()
        self.linear.add_model('linear', nn.Linear(n_feature, 1))
         # self.linear.add_model(...)

        # method3
        from collections import OrderedDict # 生成有序词典
        self.linear = nn.Sequential(OrderedDict([
                ('linear', nn.Linear(n_feature, 1)),
                # (...)
            ]))
        '''
    def forward(self, x):
        y = self.linear(x)
        return y

其中的__init__实现是可变的，可以通过nn.Sequential加入多层神经元，例如以上注释掉的部分

1.1初始化模型参数

from torch.nn import init

init.normal_(net[0].weight, mean=0.0, std=0.01)
init.constant_(net[0].bias, val=0.0)

1.2定义损失函数和优化函数

loss = nn.MSELoss()

import torch.optim as optim
optimizer = optim.SGD(net.parameters(), lr=0.03)

2.训练

大体步骤是{预测、评判、清空前一步梯度、计算本步梯度、修正权重}固定步骤

num_epochs = 3
for epoch in range(1, num_epochs + 1):
    for X, y in data_iter:
        output = net(X)
        # l/o/l/o
        l = loss(output, y.view(-1, 1))
        optimizer.zero_grad() # reset gradient, equal to net.zero_grad()
        l.backward()
        optimizer.step()
    print('epoch %d, loss: %f' % (epoch, l.item()))

查看权重

# result comparision
dense = net[0]
print(true_w, dense.weight.data)
print(true_b, dense.bias.data)