Pytorch实现简单的回归

# -*- coding: utf-8 -*-
# @Time    : 2018/1/17 18:48
# @Author  : Zhiwei Zhong
# @Site    : 回归
# @File    : Regression.py
# @Software: PyCharm
import torch
from torch.autograd import Variable
import torch.nn.functional as F
import matplotlib.pyplot as plt

# torch.unsqueeze()  把一维的数据变成二维的数据
x = torch.unsqueeze(torch.linspace(-1, 1, 100), dim=1)  # x data(tensor), shape=(100, 1)
y = x.pow(2) + 0.2 * torch.rand(x.size())  # y data(tensor), shape=(100, 1)

x, y = Variable(x), Variable(y)  # 神经网络只支持变量的形式


class Net(torch.nn.Module):
    def __init__(self, n_features, n_hidden, n_output):
        super(Net, self).__init__()     # 固定
        self.hidden = torch.nn.Linear(n_features, n_hidden)   # 输入，输出个数
        self.predict = torch.nn.Linear(n_hidden, n_output)           # 输入，输出个数

    def forward(self, x):          # 前向传递信息,x 是输入数据
        x = F.relu(self.hidden(x))
        x = self.predict(x)
        return x


net = Net(1, 10, 1)
print(net)      # 输出网络结构
optimizer = torch.optim.SGD(net.parameters(), lr=0.3)
loss_func = torch.nn.MSELoss()
plt.ion()
plt.show()
for t in range(500):
    prediction = net(x)

    loss = loss_func(prediction, y)     # 真实值在后
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()

    if t % 5 == 0:
        plt.cla()
        plt.scatter(x.data.numpy(), y.data.numpy())
        plt.plot(x.data.numpy(), prediction.data.numpy(), 'r-', lw=5)
        plt.text(0.5, 0, 'Loss=%.4f' % loss.data[0], fontdict={'size': 20, 'color': 'red'})
        plt.pause(0.1)

plt.ioff()
plt.show()