解决回归问题的神经网络

1.在pycharm中动态显示结果

import matplotlib.pyplot as plt
import torch
import numpy as np
import torch.nn as nn
x=torch.unsqueeze(torch.linspace(-np.pi,np.pi,100),dim=1)#等差数列（起点，终点，个数）并转化成二维数组
y=np.sin(x)+0.5*torch.rand(x.size())
class Net(nn.Module):
    def __init__(self):
        super(Net,self).__init__()
        self.predict=nn.Sequential(
            nn.Linear(1,20),
            nn.ReLU(),
            nn.Linear(20,20),
            nn.ReLU(),
            nn.Linear(20,1)
        )
    def forward(self,x):#前向传播过程
        prediction=self.predict(x)#将x传入网络
        return prediction

net=Net()
optimizer=torch.optim.SGD(net.parameters(),lr=0.05)
loss_func=nn.MSELoss()
plt.ion()#打开交互模式
for epoch in range(1000):
    out=net(x)
    loss=loss_func(out,y)
    optimizer.zero_grad()#清除梯度
    loss.backward()
    optimizer.step()#优化器开始优化
    if epoch % 25==0:
        plt.cla()#清除上一次绘图
        plt.scatter(x,y)
        plt.plot(x,out.data.numpy(),"r",lw=5)
        plt.text(0.5,0,f'{loss}',fontdict={'size':20,'color':'red'})
        plt.pause(0.1)
    plt.show()
plt.ioff()
plt.show()#定格最后的显示

2.在jupyter notebook中动态显示

net=Net()
optimizer=torch.optim.SGD(net.parameters(),lr=0.05)
loss_func=nn.MSELoss()
for epoch in range(1000):
    out=net(x)
    loss=loss_func(out,y)
    optimizer.zero_grad()#清除梯度
    loss.backward()
    optimizer.step()#优化器开始优化
    if epoch % 25==0:
        #print(f'迭代次数：{epoch}'+f'loss:{loss}')
        display.clear_output(wait=True)
        plt.scatter(x,y)
        plt.plot(x,out.data.numpy(),"r",lw=5)
        plt.text(0.5,0,f'{loss}',fontdict={'size':20,'color':'red'})
        plt.show()#定格最后的显示
        plt.pause(0.1)

 

jupyter notebook动态显示神经网络解决回归问题