pytorch搭建神经网络初体验

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

# In[11]:


get_ipython().run_line_magic('matplotlib', 'inline')
import torch
import random
import numpy as np 
from IPython import display
from matplotlib import pyplot as plt


# In[12]:


# 数据集
data = np.array([[1, 0, -1], [0, 1, -1],
                 [1, 1, 1], [0, 0, 1]], dtype='float32')
data = torch.from_numpy(data)

features = data[:,0:2]
labels = data[:,2]
print(features,labels)


# In[ ]:





# In[13]:


# 读取数据  简易方法
import torch.utils.data as Data
# 小批量处理
batch_size = 2
dataset = Data.TensorDataset(features,labels)
# 特征和标签组合

data_iter = Data.DataLoader(dataset,batch_size,shuffle=True)
# 随机读取小批量


for X,y in data_iter:
 print(X,y)
 break


# In[14]:

#简易搭建方法
import torch.nn as nn
net = torch.nn.Sequential(
    torch.nn.Linear(2, 2),
    torch.nn.ReLU(),
    torch.nn.Linear(2, 1)
)
print(net)


# In[15]:


nn.init.normal_(net[0].weight,mean=0 ,std= 0.01)
nn.init.constant_(net[0].bias,val=0 )

nn.init.normal_(net[2].weight,mean=0 ,std= 0.01)
nn.init.constant_(net[2].bias,val=0 )

loss = nn.MSELoss()



import torch.optim as optim

optimizer = optim.SGD(net.parameters(), lr=0.03)
print(optimizer)


num_epochs = 1000
for epoch in range(1, num_epochs + 1):
 for X, y in data_iter:
        output = net(X)
        l = loss(output, y.view(-1, 1))
        optimizer.zero_grad() # 梯度清零，等价于net.zero_grad()
        l.backward()
        optimizer.step()
 print('epoch %d, loss: %f' % (epoch, l.item()))
i=0
while i<=2:
    dense = net[i]
 print( i,'的weight',dense.weight)
 print( i,'的bias',dense.bias)
    i+=2

test = net(features)
print('最后结果是：',test)


# In[16]:


# 测试
W1 =torch.tensor([[-1.6941,  1.6941],
        [-1.0862,  1.0863]])
print(W1,W1.size())
W2 = torch.tensor([[-2.3606,  1.8405]])
print(W2,W2.size())
B1 = torch.tensor([-1.3022e-05,  1.0862e+00])
B2 = torch.tensor([-0.9993])

for i in range(0,4):
#     print(features.size())
#     print(features[i,:])
    temp = torch.mm(W1,features[i,:].view(2,1))
#     print('temp',temp.size())
    temp = temp+B1.view(2,1)
#     print('temp',temp.size())
 for j in range(len(temp)):
 if temp[j]<0:
            temp[j]=0
    ans = torch.mm(W2,temp)+B2
 print(i,'ans:',ans)


# In[21]:


import os
print(os.path.abspath('.'))

解决异或问题，只有4个数据点，样本有点少，所以收敛效果每次有差别，波动比较大，多运行几次就ok了。