pytorch实现逻辑回归训练

！ 比较简单，直接上代码

 

import torch
import torch.nn as nn
import torch.optim
from torch.autograd import Variable
from sklearn import datasets
import numpy as np

#### 读取iris数据
iris=datasets.load_iris()
x,y=iris.data,iris.target
### 前100行包括两个分类
x=x[0:100,:].astype(np.float32)
y=np.array(y[0:100],dtype=np.float32).reshape(-1,1)

#将numpy格式转为tensor
x_train=torch.from_numpy(x)
y_train=torch.from_numpy(y)

# 定义模型,输入为4维的向量，输出为1维
class logisticReg(nn.Module):
    def __init__(self):
        super(logisticReg,self).__init__()
        self.conv=nn.Sequential(
            nn.Linear(4,1),
            nn.Sigmoid()
        )

    def forward(self, x):
        y=self.conv(x)
        return y
#
model=logisticReg()
# 定义一个损失函数
criterion=nn.MSELoss()
# 定义一个优化函数
optimizer=torch.optim.SGD(model.parameters(),lr=0.01)
# 开始训练
for epoch in range(10000):
    x=Variable(x_train)
    y=Variable(y_train)
    y_pred=model(x)

    loss=criterion(y_pred,y)
    optimizer.zero_grad()
    loss.backward()
    optimizer.step()

    if epoch % 1000==0:
        print("loss:{},epoch:{}".format(loss, epoch))

 

！显示结果

loss:0.27741584181785583,epoch:0
loss:0.022335145622491837,epoch:1000
loss:0.011672687716782093,epoch:2000
loss:0.00790750328451395,epoch:3000
loss:0.005995686165988445,epoch:4000
loss:0.0048401085659861565,epoch:5000
loss:0.004065848421305418,epoch:6000
loss:0.0035105173010379076,epoch:7000
loss:0.003092461731284857,epoch:8000
loss:0.002766150049865246,epoch:9000