《PyTorch深度学习实战》第六讲

说明&＃xff1a;逻辑斯蒂回归和线性模型的明显区别是在线性模型的后面&＃xff0c;添加了激活函数(非线性变换)&＃xff0c;以增强模型的适用性

回归问题&＃xff1a;逻辑回归损失函数

二分类&＃xff1a;交叉熵损失函数

多分类&＃xff1a;Softmax分类函数

代码

import torch
import torch.nn.functional as F# 1.准备数据
x_data &＃61; torch.Tensor([[1.0], [2.0], [3.0]])
y_data &＃61; torch.Tensor([[0.], [0.], [1.]])# 2.设计模型
class LogisticRegressionModel(torch.nn.Module):def __init__(self):super(LogisticRegressionModel, self).__init__()self.linear &＃61; torch.nn.Linear(1, 1)def forward(self, x):y_pred &＃61; torch.sigmoid(self.linear(x))return y_predmodel &＃61; LogisticRegressionModel()# 3.构建损失函数和优化器
criterion &＃61; torch.nn.BCELoss(size_average&＃61;False)
optimizer &＃61; torch.optim.SGD(model.parameters(), lr&＃61;0.01)# 4.循环
for epoch in range(1000):y_pred &＃61; model(x_data)loss &＃61; criterion(y_pred, y_data)print(epoch, loss.item())optimizer.zero_grad()loss.backward()optimizer.step()# 绘图
import numpy as np
import matplotlib.pyplot as plt
import os
os.environ[&＃39;KMP_DUPLICATE_LIB_OK&＃39;] &＃61; &＃39;TRUE&＃39;
x &＃61; np.linspace(0, 10, 200)
x_t &＃61; torch.Tensor(x).view((200,1))
y_t &＃61; model(x_t)
y &＃61; y_t.data.numpy()
plt.plot(x,y)
plt.plot([0, 10], [0.5,0.5], c&＃61;&＃39;r&＃39;)
plt.xlabel(&＃39;Hours&＃39;)
plt.ylabel(&＃39;Probability of Pass&＃39;)
plt.grid()
plt.show()