python构建深度神经网络(续)


Posted in Python onMarch 10, 2018

这篇文章在前一篇文章:python构建深度神经网络(DNN)的基础上,添加了一下几个内容:

1) 正则化项

2) 调出中间损失函数的输出

3) 构建了交叉损失函数

4) 将训练好的网络进行保存,并调用用来测试新数据

1  数据预处理

#!/usr/bin/env python 
# -*- coding: utf-8 -*- 
# @Time : 2017-03-12 15:11 
# @Author : CC 
# @File : net_load_data.py 
 
from numpy import * 
import numpy as np 
import cPickle 
def load_data(): 
 """载入解压后的数据,并读取""" 
 with open('data/mnist_pkl/mnist.pkl','rb') as f: 
  try: 
   train_data,validation_data,test_data = cPickle.load(f) 
   print " the file open sucessfully" 
   # print train_data[0].shape #(50000,784) 
   # print train_data[1].shape #(50000,) 
   return (train_data,validation_data,test_data) 
  except EOFError: 
   print 'the file open error' 
   return None 
 
def data_transform(): 
 """将数据转化为计算格式""" 
 t_d,va_d,te_d = load_data() 
 # print t_d[0].shape # (50000,784) 
 # print te_d[0].shape # (10000,784) 
 # print va_d[0].shape # (10000,784) 
 # n1 = [np.reshape(x,784,1) for x in t_d[0]] # 将5万个数据分别逐个取出化成(784,1),逐个排列 
 n = [np.reshape(x, (784, 1)) for x in t_d[0]] # 将5万个数据分别逐个取出化成(784,1),逐个排列 
 # print 'n1',n1[0].shape 
 # print 'n',n[0].shape 
 m = [vectors(y) for y in t_d[1]] # 将5万标签(50000,1)化为(10,50000) 
 train_data = zip(n,m) # 将数据与标签打包成元组形式 
 n = [np.reshape(x, (784, 1)) for x in va_d[0]] # 将5万个数据分别逐个取出化成(784,1),排列 
 validation_data = zip(n,va_d[1]) # 没有将标签数据矢量化 
 n = [np.reshape(x, (784, 1)) for x in te_d[0]] # 将5万个数据分别逐个取出化成(784,1),排列 
 test_data = zip(n, te_d[1]) # 没有将标签数据矢量化 
 # print train_data[0][0].shape #(784,) 
 # print "len(train_data[0])",len(train_data[0]) #2 
 # print "len(train_data[100])",len(train_data[100]) #2 
 # print "len(train_data[0][0])", len(train_data[0][0]) #784 
 # print "train_data[0][0].shape", train_data[0][0].shape #(784,1) 
 # print "len(train_data)", len(train_data) #50000 
 # print train_data[0][1].shape #(10,1) 
 # print test_data[0][1] # 7 
 return (train_data,validation_data,test_data) 
def vectors(y): 
 "赋予标签" 
 label = np.zeros((10,1)) 
 label[y] = 1.0 #浮点计算 
 return label

2 网络定义和训练

#!/usr/bin/env python 
# -*- coding: utf-8 -*- 
# @Time : 2017-03-28 10:18 
# @Author : CC 
# @File : net_network2.py 
 
from numpy import * 
import numpy as np 
import operator 
import json 
# import sys 
 
class QuadraticCost(): 
 """定义二次代价函数类的方法""" 
 @staticmethod 
 def fn(a,y): 
  cost = 0.5*np.linalg.norm(a-y)**2 
  return cost 
 @staticmethod 
 def delta(z,a,y): 
  delta = (a-y)*sig_derivate(z) 
  return delta 
 
class CrossEntroyCost(): 
 """定义交叉熵函数类的方法""" 
 @staticmethod 
 def fn(a, y): 
  cost = np.sum(np.nan_to_num(-y*np.log(a)-(1-y)*np.log(1-a))) # not a number---0, inf---larger number 
  return cost 
 @staticmethod 
 def delta(z, a, y): 
  delta = (a - y) 
  return delta 
 
class Network(object): 
 """定义网络结构和方法""" 
 def __init__(self,sizes,cost): 
  self.num_layer = len(sizes) 
  self.sizes = sizes 
  self.cost = cost 
  # print "self.cost.__name__:",self.cost.__name__ # CrossEntropyCost 
  self.default_weight_initializer() 
 def default_weight_initializer(self): 
  """权值初始化""" 
  self.bias = [np.random.rand(x, 1) for x in self.sizes[1:]] 
  self.weight = [np.random.randn(y, x)/float(np.sqrt(x)) for (x, y) in zip(self.sizes[:-1], self.sizes[1:])] 
 
 def large_weight_initializer(self): 
  """权值另一种初始化""" 
  self.bias = [np.random.rand(x, 1) for x in self.sizes[1:]] 
  self.weight = [np.random.randn(y, x) for x, y in zip(self.sizes[:-1], self.sizes[1:])] 
 def forward(self,a): 
  """forward the network""" 
  for w,b in zip(self.weight,self.bias): 
   a=sigmoid(np.dot(w,a)+b) 
  return a 
 
 def SGD(self,train_data,min_batch_size,epochs,eta,test_data=False, 
   lambd = 0, 
   monitor_train_cost = False, 
   monitor_train_accuracy = False, 
   monitor_test_cost=False, 
   monitor_test_accuracy=False 
   ): 
  """1)Set the train_data,shuffle; 
   2) loop the epoches, 
   3) set the min_batches,and rule of update""" 
  if test_data: n_test=len(test_data) 
  n = len(train_data) 
  for i in xrange(epochs): 
   random.shuffle(train_data) 
   min_batches = [train_data[k:k+min_batch_size] for k in xrange(0,n,min_batch_size)] 
 
   for min_batch in min_batches: # 每次提取一个批次的样本 
    self.update_minbatch_parameter(min_batch,eta,lambd,n) 
   train_cost = [] 
   if monitor_train_cost: 
    cost1 = self.total_cost(train_data,lambd,cont=False) 
    train_cost.append(cost1) 
    print "epoche {0},train_cost: {1}".format(i,cost1) 
   if monitor_train_accuracy: 
    accuracy = self.accuracy(train_data,cont=True) 
    train_cost.append(accuracy) 
    print "epoche {0}/{1},train_accuracy: {2}".format(i,epochs,accuracy) 
   test_cost = [] 
   if monitor_test_cost: 
    cost1 = self.total_cost(test_data,lambd) 
    test_cost.append(cost1) 
    print "epoche {0},test_cost: {1}".format(i,cost1) 
   test_accuracy = [] 
   if monitor_test_accuracy: 
    accuracy = self.accuracy(test_data) 
    test_cost.append(accuracy) 
    print "epoche:{0}/{1},test_accuracy:{2}".format(i,epochs,accuracy) 
  self.save(filename= "net_save") #保存网络网络参数 
 
 def total_cost(self,train_data,lambd,cont=True): 
  cost1 = 0.0 
  for x,y in train_data: 
   a = self.forward(x) 
   if cont: y = vectors(y) #将测试样本标签化为矩阵 
   cost1 += (self.cost).fn(a,y)/len(train_data) 
  cost1 += lambd/len(train_data)*np.sum(np.linalg.norm(weight)**2 for weight in self.weight) #加上权值项 
  return cost1 
 def accuracy(self,train_data,cont=False): 
  if cont: 
   output1 = [(np.argmax(self.forward(x)),np.argmax(y)) for (x,y) in train_data] 
  else: 
   output1 = [(np.argmax(self.forward(x)), y) for (x, y) in train_data] 
  return sum(int(out1 == y) for (out1, y) in output1) 
 def update_minbatch_parameter(self,min_batch, eta,lambd,n): 
  """1) determine the weight and bias 
   2) calculate the the delta 
   3) update the data """ 
  able_b = [np.zeros(b.shape) for b in self.bias] 
  able_w=[np.zeros(w.shape) for w in self.weight] 
  for x,y in min_batch: #每次只取一个样本? 
   deltab,deltaw = self.backprop(x,y) 
   able_b =[a_b+dab for a_b, dab in zip(able_b,deltab)] #实际上对dw,db做批次累加,最后小批次取平均 
   able_w = [a_w + daw for a_w, daw in zip(able_w, deltaw)] 
  self.weight = [weight - eta * (dw) / len(min_batch)- eta*(lambd*weight)/n for weight, dw in zip(self.weight,able_w) ] 
  #增加正则化项:eta*lambda/m *weight 
  self.bias = [bias - eta * db / len(min_batch) for bias, db in zip(self.bias, able_b)] 
 
 def backprop(self,x,y): 
  """" 1) clacu the forward value 
   2) calcu the delta: delta =(y-f(z)); deltak = delta*w(k)*fz(k-1)' 
   3) clacu the delta in every layer: deltab=delta; deltaw=delta*fz(k-1)""" 
  deltab = [np.zeros(b.shape) for b in self.bias] 
  deltaw = [np.zeros(w.shape) for w in self.weight] 
  zs = [] 
  activate = x 
  activates = [x] 
  for w,b in zip(self.weight,self.bias): 
   z =np.dot(w, activate) +b 
   zs.append(z) 
   activate = sigmoid(z) 
   activates.append(activate) 
   # backprop 
  delta = self.cost.delta(zs[-1],activates[-1],y) #调用不同代价函数的方法求梯度 
  deltab[-1] = delta 
  deltaw[-1] = np.dot(delta ,activates[-2].transpose()) 
  for i in xrange(2,self.num_layer): 
   z = zs[-i] 
   delta = np.dot(self.weight[-i+1].transpose(),delta)* sig_derivate(z) 
   deltab[-i] = delta 
   deltaw[-i] = np.dot(delta,activates[-i-1].transpose()) 
  return (deltab,deltaw) 
 
 def save(self,filename): 
  """将训练好的网络采用json(java script object notation)将对象保存成字符串保存,用于生产部署 
  encoder=json.dumps(data) 
  python 原始类型(没有数组类型)向 json 类型的转化对照表: 
   python    json 
   dict    object 
  list/tuple   arrary 
  int/long/float  number 
  .tolist() 将数组转化为列表 
  >>> a = np.array([[1, 2], [3, 4]]) 
  >>> list(a) 
  [array([1, 2]), array([3, 4])] 
  >>> a.tolist() 
  [[1, 2], [3, 4]] 
  """ 
  data = {"sizes": self.sizes,"weight": [weight.tolist() for weight in self.weight], 
    "bias": ([bias.tolist() for bias in self.bias]), 
    "cost": str(self.cost.__name__)} 
  # 保存网络训练好的权值,偏置,交叉熵参数。 
  f = open(filename, "w") 
  json.dump(data,f) 
  f.close() 
 
def load_net(filename): 
 """采用data=json.load(json.dumps(data))进行解码, 
 decoder = json.load(encoder) 
 编码后和解码后键不会按照原始data的键顺序排列,但每个键对应的值不会变 
 载入训练好的网络用于测试""" 
 f = open(filename,"r") 
 data = json.load(f) 
 f.close() 
 # print "data[cost]", getattr(sys.modules[__name__], data["cost"])#获得属性__main__.CrossEntropyCost 
 # print "data[cost]", data["cost"], data["sizes"] 
 net = Network(data["sizes"], cost=data["cost"]) #网络初始化 
 net.weight = [np.array(w) for w in data["weight"]] #赋予训练好的权值,并将list--->array 
 net.bias = [np.array(b) for b in data["bias"]] 
 return net 
 
def sig_derivate(z): 
 """derivate sigmoid""" 
 return sigmoid(z) * (1-sigmoid(z)) 
 
def sigmoid(x): 
 sigm=1.0/(1.0+exp(-x)) 
 return sigm 
 
def vectors(y): 
 """赋予标签""" 
 label = np.zeros((10,1)) 
 label[y] = 1.0 #浮点计算 
 return label

3) 网络测试

#!/usr/bin/env python 
# -*- coding: utf-8 -*- 
# @Time : 2017-03-12 15:24 
# @Author : CC 
# @File : net_test.py 
 
import net_load_data 
# net_load_data.load_data() 
train_data,validation_data,test_data = net_load_data.data_transform() 
 
import net_network2 as net 
cost = net.QuadraticCost 
cost = net.CrossEntroyCost 
lambd = 0 
net1 = net.Network([784,50,10],cost) 
min_batch_size = 30 
eta = 3.0 
epoches = 2 
net1.SGD(train_data,min_batch_size,epoches,eta,test_data, 
   lambd, 
   monitor_train_cost=True, 
   monitor_train_accuracy=True, 
   monitor_test_cost=True, 
   monitor_test_accuracy=True 
   ) 
print "complete"

4 调用训练好的网络进行测试

#!/usr/bin/env python 
# -*- coding: utf-8 -*- 
# @Time : 2017-03-28 17:27 
# @Author : CC 
# @File : forward_test.py 
 
import numpy as np 
# 对训练好的网络直接进行调用,并用测试样本进行测试 
import net_load_data #导入测试数据 
import net_network2 as net 
train_data,validation_data,test_data = net_load_data.data_transform() 
net = net.load_net(filename= "net_save")  #导入网络 
output = [(np.argmax(net.forward(x)),y) for (x,y) in test_data] #测试 
print sum(int(y1 == y2) for (y1,y2) in output)  #输出最终值

以上就是本文的全部内容,希望对大家的学习有所帮助,也希望大家多多支持三水点靠木。

Python 相关文章推荐
python遍历类中所有成员的方法
Mar 18 Python
python使用Flask框架获取用户IP地址的方法
Mar 21 Python
ubuntu系统下 python链接mysql数据库的方法
Jan 09 Python
Python处理命令行参数模块optpars用法实例分析
May 31 Python
python 检查文件mime类型的方法
Dec 08 Python
Scrapy框架爬取西刺代理网免费高匿代理的实现代码
Feb 22 Python
如何使用pyinstaller打包32位的exe程序
May 26 Python
Django后端接收嵌套Json数据及解析详解
Jul 17 Python
详解基于python的多张不同宽高图片拼接成大图
Sep 26 Python
python3+selenium获取页面加载的所有静态资源文件链接操作
May 04 Python
Python命名空间namespace及作用域原理解析
Jun 05 Python
Python中如何处理常见报错
Jan 18 Python
python构建深度神经网络(DNN)
Mar 10 #Python
Python使用numpy实现BP神经网络
Mar 10 #Python
python实现日常记账本小程序
Mar 10 #Python
python实现简单神经网络算法
Mar 10 #Python
TensorFlow saver指定变量的存取
Mar 10 #Python
TensorFLow用Saver保存和恢复变量
Mar 10 #Python
tensorflow创建变量以及根据名称查找变量
Mar 10 #Python
You might like
PHP新手上路(十)
2006/10/09 PHP
WindowsXP中快速配置Apache+PHP5+Mysql
2008/06/05 PHP
php select,radio和checkbox默认选择的实现方法
2010/05/15 PHP
ThinkPHP提交表单时默认自动转义的解决方法
2014/11/25 PHP
详解Yii2.0 rules验证规则集合
2017/03/21 PHP
php利用ZipArchive类操作文件的实例
2020/01/21 PHP
clientX,pageX,offsetX,x,layerX,screenX,offsetLeft区别分析
2010/03/12 Javascript
使用js操作cookie的一点小收获分享
2013/09/03 Javascript
IE8下String的Trim()方法失效的解决方法
2013/11/08 Javascript
jquery 3D 标签云示例代码
2014/06/12 Javascript
浅析Node.js查找字符串功能
2014/09/03 Javascript
原生JS实现美图瀑布流布局赏析
2015/09/07 Javascript
js查看一个函数的执行时间实例代码
2015/09/12 Javascript
javascript单页面手势滑屏切换原理详解
2016/03/21 Javascript
AngularJS自定义服务与fliter的混合使用
2016/11/24 Javascript
图片懒加载插件实例分享(含解析)
2017/01/09 Javascript
微信小程序基于本地缓存实现点赞功能的方法
2017/12/18 Javascript
使用mock.js随机数据和使用express输出json接口的实现方法
2018/01/07 Javascript
详解JS中统计函数执行次数与执行时间
2018/09/04 Javascript
重学JS 系列:聊聊继承(推荐)
2019/04/11 Javascript
微信公众平台 发送模板消息(Java接口开发)
2019/04/17 Javascript
Python牛刀小试密码爆破
2011/02/03 Python
python入门之语句(if语句、while语句、for语句)
2015/01/19 Python
python 执行shell命令并将结果保存的实例
2018/05/11 Python
Python计算一个给定时间点前一个月和后一个月第一天的方法
2018/05/29 Python
Python代码打开本地.mp4格式文件的方法
2019/01/03 Python
详解js文件通过python访问数据库方法
2019/03/03 Python
python mongo 向数据中的数组类型新增数据操作
2020/12/05 Python
全球性的在线婚纱礼服工厂:27dress.com
2019/03/21 全球购物
专科毕业生就业推荐信
2013/11/01 职场文书
质检的岗位职责
2013/11/17 职场文书
夜班门卫岗位职责
2013/12/09 职场文书
校长师德表现自我评价
2015/03/05 职场文书
大学生创业,为什么都会选择快餐饮?
2019/08/08 职场文书
微信小程序scroll-view不能左右滑动问题的解决方法
2021/07/09 Javascript
python神经网络Xception模型
2022/05/06 Python