Pytorch入门＞简单的线性回归2

看李沐老师课程实现的一个简单的线性回归，在之前的基础上主要开始利用官方的库创建数据库和搭建模型。

线性回归+MSELoss+SGD优化。

#!/usr/bin/env python
# coding: utf-8

# In[3]:


import torch
import numpy as np
from torch.utils import data
import matplotlib.pyplot as plt


# In[4]:


def generate_data(w,b,num):
    x = torch.normal(0,1, (num,len(w)))
    y = torch.matmul(x,w)+b
    y+=torch.normal(0,0.01,y.shape)
    return x,y.reshape((-1,1))


# In[5]:


true_w = torch.tensor([2.,-3.])
true_b = 0.2
feature, label =generate_data(true_w,true_b,200)


# In[6]:


def load_array(data_arrays,batch_size,is_train=True):#pytorch数据迭代器
    dataset = data.TensorDataset(*data_arrays)
    return data.DataLoader(dataset,batch_size,shuffle=is_train)


# In[7]:


batch_size = 10
data_iter = load_array((feature,label),batch_size)

next(iter(data_iter)) #python中的iter


# In[8]:


#模型的定义
from torch import nn
net = nn.Sequential(nn.Linear(2,1))


# In[9]:


#自定义模型初始化参数
print(net[0].weight)
print(net[0].weight.data.normal_(0,0.01))#normal为正太分布
print(net[0].bias)


# In[10]:


loss =nn.MSELoss()
trainer = torch.optim.SGD(net.parameters(),lr = 0.03)#SGD至少需要传入2个参数：模型的参数和学习率


# In[14]:


num_epoch =3
for epoch in range(num_epoch):
    for x,y in data_iter:
        l =loss(net(x),y)
        trainer.zero_grad()
        l.backward()
        trainer.step()
        
    l = loss(net(feature),label)
    print(f'epoch {epoch+1},loss{l:f}')
          


# In[ ]: