Merge pull request BaiduOSS#6 from tanzhongyibidu/master

update to fix style problem

Author: tanzhongyi

LICENSE

@@ -1,4 +1,4 @@
-Copyright (c) 2016 PaddlePaddle Authors. All Rights Reserved
+Copyright (c) 2018 Baidu Inc. All Rights Reserved
 
  Apache License
  Version 2.0, January 2004

README.md

@@ -2,13 +2,13 @@
 
 
 it includes
-* lesson1 python basic
-* lesson2 paddle basic
-* lesson3 logistic classification with numpy and paddle
-* lesson4 shallow network with numpy and paddle
-* lesson5 deep network with numpy and paddle
-* lesson6 mnist with paddle
-* lesson7 recommend system with paddle
-* lesson8 paddlecloud
-* lesson9 ctr
-* lesson10 tuning
+* lesson1 Python basic
+* lesson2 PaddlePaddle basic
+* lesson3 Logistic classification with numpy and paddle
+* lesson4 Shallow network with numpy and paddle
+* lesson5 Deep network with numpy and paddle
+* lesson6 Mnist with paddle
+* lesson7 Recommend system with paddle
+* lesson8 PaddlePaddleCloud
+* lesson9 CTR
+* lesson10 Parameter tuning

lesson1/array_module.py

@@ -1,27 +1,36 @@
-#!/usr/bin/env python2.7
-# -*- encoding:utf-8 -*-
+#!/usr/bin/env python
+# -*- coding:utf-8 -*-
+
 """
 Array 基础操作脚本
 Created on 2017-11-16
-@author: denglelai@baidu.com
-@copyright: www.baidu.com
+author: denglelai
 """
+
 import numpy as np
 
 
-# a是python中的list类型
-a = [1, 2, 3, 4]
-# 数组化之后b的类型变为 array
-b = np.array(a)
-# a的类型 <type 'list'>
-print "原始的类型为：" + str(type(a))
-# b的类型 <type 'numpy.ndarray'>
-print "数组化之后的类型为：" + str(type(b))
-# shape参数表示array的大小，这里是(4,)
-print "Array的大小为：" + str(b.shape)
-# 调用argmax()函数可以求得array中的最大值的索引，这里是3
-print "Array中最大元素索引为：" + str(b.argmax())
-# 调用max()函数可以求得array中的最大值，这里是4
-print "Array中最大元素值为：" + str(b.max())
-# 调用mean()函数可以求得array中的平均值，这里是2.5
-print "Array中元素平均值为：" + str(b.mean())
+def main():
+ """
+ show np.array related operations
+ """
+ # a是python中的list类型
+ a = [1, 2, 3, 4]
+ # 数组化之后b的类型变为 array
+ b = np.array(a)
+ # a的类型 <type 'list'>
+ print "原始的类型为：" + str(type(a))
+ # b的类型 <type 'numpy.ndarray'>
+ print "数组化之后的类型为：" + str(type(b))
+ # shape参数表示array的大小，这里是(4,)
+ print "Array的大小为：" + str(b.shape)
+ # 调用argmax()函数可以求得array中的最大值的索引，这里是3
+ print "Array中最大元素索引为：" + str(b.argmax())
+ # 调用max()函数可以求得array中的最大值，这里是4
+ print "Array中最大元素值为：" + str(b.max())
+ # 调用mean()函数可以求得array中的平均值，这里是2.5
+ print "Array中元素平均值为：" + str(b.mean())
+
+
+if __name__ == '__main__':
+ main()

lesson1/basic_calculate.py

@@ -1,50 +1,59 @@
-#!/usr/bin/env python2.7
-# -*- encoding:utf-8 -*-
+#!/usr/bin/env python
+# -*- coding:utf-8 -*-
+
 """
 Numpy中实现基本数学计算脚本
 Created on 2017-11-16
-@author: denglelai@baidu.com
-@copyright: www.baidu.com
+author: denglelai
 """
 import numpy as np
 
 
-# 绝对值，1
-a = np.abs(-1)
-print "-1的绝对值为:" + str(a)
+def main():
+ """
+ run basic operations of numpy
+ """
+
+ # 绝对值，1
+ a_variable = np.abs(-1)
+ print "-1的绝对值为:" + str(a_variable)
+
+ # sin函数，1.0
+ a_variable = np.sin(np.pi / 2)
+ print "pi/2的正弦值为:" + str(a_variable)
+
+ # tanh逆函数，0.500001071578
+ a_variable = np.arctanh(0.462118)
+ print "tanh(0.462118)值为:" + str(a_variable)
 
-# sin函数，1.0
-b = np.sin(np.pi / 2)
-print "pi/2的正弦值为:" + str(b)
+ # e为底的指数函数，20.0855369232
+ a_variable = np.exp(3)
+ print "e的3次方值为:" + str(a_variable)
 
-# tanh逆函数，0.50000107157840523
-c = np.arctanh(0.462118)
-print "tanh(0.462118)值为:" + str(c)
+ # 2的3次方，8
+ a_variable = np.power(2, 3)
+ print "2的3次方值为:" + str(a_variable)
 
-# e为底的指数函数，20.085536923187668
-d = np.exp(3)
-print "e的3次方值为:" + str(d)
+ # 点积，1*3+2*4=11
+ a_variable = np.dot([1, 2], [3, 4])
+ print "向量[1. 2]与向量[3. 4]点乘值为:" + str(a_variable)
 
-# 2的3次方，8
-f = np.power(2, 3)
-print "2的3次方值为:" + str(f)
+ # 开方，5
+ a_variable = np.sqrt(25)
+ print "25的2次方根值为:" + str(a_variable)
 
-# 点积，1*3+2*4=11
-g = np.dot([1, 2], [3, 4])
-print "向量[1. 2]与向量[3. 4]点乘值为:" + str(g)
+ # 求和，10
+ a_variable = np.sum([1, 2, 3, 4])
+ print "对[1, 2, 3, 4]中元素求和结果为:" + str(a_variable)
 
-# 开方，5
-h = np.sqrt(25)
-print "25的2次方根值为:" + str(h)
+ # 平均值，5.5
+ a_variable = np.mean([4, 5, 6, 7])
+ print "对[1, 2, 3, 4]中元素求平均结果为:" + str(a_variable)
 
-# 求和，10
-l = np.sum([1, 2, 3, 4])
-print "对[1, 2, 3, 4]中元素求和结果为:" + str(l)
+ # 标准差，0.968245836552
+ a_variable = np.std([1, 2, 3, 2, 1, 3, 2, 0])
+ print "对[1, 2, 3, 2, 1, 3, 2, 0]中元素求标准差结果为:" + str(a_variable)
 
-# 平均值，5.5
-m = np.mean([4, 5, 6, 7])
-print "对[1, 2, 3, 4]中元素求平均结果为:" + str(m)
 
-# 标准差，0.96824583655185426
-p = np.std([1, 2, 3, 2, 1, 3, 2, 0])
-print "对[1, 2, 3, 2, 1, 3, 2, 0]中元素求标准差结果为:" + str(p)
+if __name__ == '__main__':
+ main()

lesson1/broadcast.py

@@ -1,45 +1,50 @@
-#!/usr/bin/env python2.7
-# -*- encoding:utf-8 -*-
+#!/usr/bin/env python
+# -*- coding:utf-8 -*-
+
 """
 Numpy中实现广播机制说明脚本
 Created on 2017-11-16
-@author: denglelai@baidu.com
-@copyright: www.baidu.com
+author: denglelai
 """
 import numpy as np
 
 
-a = np.array([
- [1, 2, 3],
- [4, 5, 6]
-])
-
-b = np.array([
- [1, 2, 3],
- [1, 2, 3]
-])
-
-'''
-维度一样的array，对位计算
-array([[2, 4, 6],
- [5, 7, 9]])
-'''
-print "相同维度array, 进行对位运算, 结果为：\n" + str(a + b)
-
-c = np.array([
- [1, 2, 3],
- [4, 5, 6],
- [7, 8, 9],
- [10, 11, 12]
-])
-d = np.array([2, 2, 2])
-
-'''
-广播机制让计算的表达式保持简洁
-d和c的每一行分别进行运算
-array([[ 3, 4, 5],
- [ 6, 7, 8],
- [ 9, 10, 11],
- [12, 13, 14]])
-'''
-print "广播机制下, c和d进行每一行分别计算, 结果为：\n" + str(c + d)
+def main():
+ """
+ show broadcast operation in numpy
+ """
+ array_a = np.array([
+ [1, 2, 3],
+ [4, 5, 6]
+ ])
+
+ array_b = np.array([
+ [1, 2, 3],
+ [1, 2, 3]
+ ])
+
+ # 维度一样的array，对位计算
+ # array([[2, 4, 6],
+ # [5, 7, 9]])
+
+ print "相同维度array, 进行对位运算, 结果为：\n" + str(array_a + array_b)
+
+ array_c = np.array([
+ [1, 2, 3],
+ [4, 5, 6],
+ [7, 8, 9],
+ [10, 11, 12]
+ ])
+
+ array_d = np.array([2, 2, 2])
+ # 广播机制让计算的表达式保持简洁
+ # array_d和array_c的每一行分别进行运算
+ # array([[ 3, 4, 5],
+ # [ 6, 7, 8],
+ # [ 9, 10, 11],
+ # [12, 13, 14]])
+ print "广播机制下, c和d进行每一行分别计算, 结果为：\n" + str(array_c + array_d)
+
+
+if __name__ == '__main__':
+ main()

lesson1/compare_efficiency.py

@@ -1,31 +1,40 @@
-#!/usr/bin/env python2.7
-# -*- encoding:utf-8 -*-
+#!/usr/bin/env python
+# -*- coding:utf-8 -*-
+
 """
 Numpy中实现向量化与非向量化计算效率比较脚本
 Created on 2017-11-16
-@author: denglelai@baidu.com
-@copyright: www.baidu.com
+author: denglelai
 """
-import numpy as np
 import time
+import numpy as np
+
+
+def main():
+ """
+ show two version of vector operation
+ """
+
+ # 初始化两个1000000维的随机向量v1,v2用于矩阵相乘计算
+ vector_1 = np.random.rand(1000000)
+ vector_2 = np.random.rand(1000000)
+ result = 0
+
+ # 矩阵相乘-非向量化版本
+ tic = time.time()
+ for i in range(1000000):
+ result = result + vector_1[i] * vector_2[i]
+ toc = time.time()
+ print "非向量化-计算结果：" + str(result)
+ print "非向量化-计算时间：" + str((toc - tic) * 1000) + "ms" + "\n"
+
+ # 矩阵相乘-向量化版本
+ tic = time.time()
+ result = np.dot(vector_1, vector_2)
+ toc = time.time()
+ print "向量化-计算结果：" + str(result)
+ print "向量化-计算时间：" + str((toc - tic) * 1000)+"ms"
 
 
-#初始化两个1000000维的随机向量v1,v2用于矩阵相乘计算
-v1 = np.random.rand(1000000)
-v2 = np.random.rand(1000000)
-v = 0
-
-#矩阵相乘-非向量化版本
-tic = time.time()
-for i in range(1000000):
- v = v + v1[i] * v2[i]
-toc = time.time()
-print "非向量化-计算结果：" + str(v) + "ms"
-print "非向量化-计算时间：" + str((toc - tic) * 1000) + "ms" + "\n"
-
-#矩阵相乘-向量化版本
-tic = time.time()
-v = np.dot(v1, v2)
-toc = time.time()
-print "向量化-计算结果：" + str(v) + "ms"
-print "向量化-计算时间：" + str((toc - tic) * 1000)+"ms"
+if __name__ == '__main__':
+ main()