pytorch实现函数拟合

导入必要模块随机生成数据

import random
import numpy as np
import matplotlib.pyplot as plt
import torch
import torch.nn as nn
import torch.nn.functional as F

x_train_list = []
y_train_list = []
for i in range(1, 50):
    x = i*random.choice([0.7,0.8,0.9])
    y = i*random.choice([0.4,0.5,0.8,0.9])
    x_train_list.append(["%.2f" % x])
    y_train_list.append(["%.2f" % y])

x_train = np.array(x_train_list, dtype=np.float32) #将数据列表转为np.array
y_train = np.array(y_train_list, dtype=np.float32)

绘图以及将np.array转为tensor

plt.scatter(x_train,y_train)
plt.show()
x_train = torch.from_numpy(x_train)
y_train = torch.from_numpy(y_train)

模型构建以及 训练数据

 

class LinearRegression(nn.Module):
  def __init__(self):
    super(LinearRegression,self).__init__()
    self.linear = nn.Linear(1,10)
    self.linear1 = nn.Linear(10,10)
    self.linear1_1 = nn.Linear(10,100)
    self.linear2 = nn.Linear(100,1)
  def forward(self,x):
    
    x = self.linear(x)
    for _ in range(3):
      x = nn.functional.relu(self.linear1(x))
    x = nn.functional.relu(x)
    x = nn.functional.relu(self.linear1_1(x))
    x = self.linear2(x)
    return x

# if torch.cuda.is_available():
#   model = LinearRegression().cuda()
# else:
model = LinearRegression()
criterion = nn.MSELoss()
import torch.optim as optim
optimizer = optim.SGD(model.parameters(),lr=1e-3)
num_epochs = 8000
for epoch in range(num_epochs):
  # if torch.cuda.is_available():
  #   input = torch.autograd.Variable(x_train).cuda()
  #   target = torch.autograd.Variable(y_train).cuda()
  # else:
  input = torch.autograd.Variable(x_train)
  target = torch.autograd.Variable(y_train)
  out = model(input)
  loss = criterion(out, target)
  optimizer.zero_grad() #清除上一梯度
  loss.backward() #梯度计算
  
  optimizer.step()#梯度优化
  if (epoch+1) % 200 == 0:
    predict = model(torch.autograd.Variable(x_train))
    predict = predict.data.numpy()
    plt.cla()
    plt.plot(x_train.numpy(),y_train.numpy(),'ro',label="original data")
    plt.plot(x_train.numpy(),predict,label='Fitting Line')
    plt.text(20,0,'Epoch[{}/{}],loss:{:.4f}'.format(epoch, num_epochs,loss.item()),fontdict={'size':12,'color':'red'})
    plt.pause(0.1)
plt.ioff()
plt.show()
    #print('Epoch[{}/{}],loss:{:.4f}'.format(epoch, num_epochs,loss.item()))

拟合结果图