(1)一维线性回归理论基础
(2)代码实现
# -*- coding: utf-8 -*-
"""
Created on Sat Apr 27 10:10:04 2019
@author: hp
"""
import torch
import torch.nn as nn #引入torch.nn并指定别名
import numpy as np
import matplotlib.pyplot as plt
# 定义参数,迭代次数为60次
input_size = 1
output_size = 1
num_epochs = 60
learning_rate = 0.001
# 数据集
x_train = np.array([[3.3], [4.4], [5.5], [6.71], [6.93], [4.168],
[9.779], [6.182], [7.59], [2.167], [7.042],
[10.791], [5.313], [7.997], [3.1]], dtype=np.float32)
y_train = np.array([[1.7], [2.76], [2.09], [3.19], [1.694], [1.573],
[3.366], [2.596], [2.53], [1.221], [2.827],
[3.465], [1.65], [2.904], [1.3]], dtype=np.float32)
# 线性回归模型
model = nn.Linear(input_size, output_size)
# 损失函数和优化器的选择
criterion = nn.MSELoss()
optimizer = torch.optim.SGD(model.parameters(), lr=learning_rate)
# Train the model
for epoch in range(num_epochs):
# Convert numpy arrays to torch tensors
inputs = torch.from_numpy(x_train)
targets = torch.from_numpy(y_train)
# Forward pass 进行前向传播
outputs = model(inputs)
loss = criterion(outputs, targets)
# Backward and optimize 反向传播
optimizer.zero_grad() #先将梯度归零
loss.backward()
optimizer.step()
if (epoch+1) % 5 == 0:
print ('Epoch [{}/{}], Loss: {:.4f}'.format(epoch+1, num_epochs, loss.item()))
# Plot the graph 画图
predicted = model(torch.from_numpy(x_train)).detach().numpy()
plt.plot(x_train, y_train, 'ro', label='Original data')
plt.plot(x_train, predicted, label='Fitted line')
plt.legend()
plt.show()
# Save the model checkpoint 保存模型
torch.save(model.state_dict(), 'model.ckpt')
(3)结果
Epoch [5/60], Loss: 0.5609
Epoch [10/60], Loss: 0.4818
Epoch [15/60], Loss: 0.4493
Epoch [20/60], Loss: 0.4358
Epoch [25/60], Loss: 0.4299
Epoch [30/60], Loss: 0.4271
Epoch [35/60], Loss: 0.4256
Epoch [40/60], Loss: 0.4246
Epoch [45/60], Loss: 0.4238
Epoch [50/60], Loss: 0.4231
Epoch [55/60], Loss: 0.4224
Epoch [60/60], Loss: 0.4218