自定义博客皮肤VIP专享

*博客头图:

格式为PNG、JPG,宽度*高度大于1920*100像素,不超过2MB,主视觉建议放在右侧,请参照线上博客头图

请上传大于1920*100像素的图片!

博客底图:

图片格式为PNG、JPG,不超过1MB,可上下左右平铺至整个背景

栏目图:

图片格式为PNG、JPG,图片宽度*高度为300*38像素,不超过0.5MB

主标题颜色:

RGB颜色,例如:#AFAFAF

Hover:

RGB颜色,例如:#AFAFAF

副标题颜色:

RGB颜色,例如:#AFAFAF

自定义博客皮肤

-+
  • 博客(41)
  • 收藏
  • 关注

转载 3.14 Simplify Path

http://www.cnblogs.com/TenosDoIt/p/3465328.htmlGiven an absolute path for a file (Unix-style), simplify it.For example, path = "/home/", => "/home" path = "/a/./b/../../c/", => "/c"

2015-05-29 15:42:09 353

原创 size_t 与auto 用法

size_t 类型定义在cstddef头文件中,该文件是C标准库的头文件stddef.h的C++版。它是一个与机器相关的unsigned类型,其大小足以保证存储内存中对象的大小。例如:在用下标访问元素时,vector使用vector::size_type作为下标类型,而数组下标的正确类型则是size_t。vector使用的下标实际也是size_t,源码是typedef size_t siz

2015-05-29 14:21:27 1241

转载 深度学习研究理解:Very Deep Convolutional Networks for Large-Scale Image Recognition

原文:http://blog.csdn.net/whiteinblue/article/details/43560491本文是牛津大学 visual geometry group(VGG)Karen Simonyan 和Andrew Zisserman 于14年撰写的论文,主要探讨了深度对于网络的重要性;并建立了一个19层的深度网络获得了很好的结果;在ILSVRC上定位第一,分类第二。

2015-05-29 11:58:08 837

翻译 alexnet

接下来本文将一步步对该网络配置结构中各个层进行详细的解读(训练阶段):1. conv1阶段DFD(data flow diagram):2. conv2阶段DFD(data flow diagram):3. conv3阶段DFD(data flow diagram):4. conv4阶段DFD(data flow diagram):5. co

2015-05-29 11:42:02 6177

翻译 caffe loss

损失函数,一般由两项组成,一项是loss term,另外一项是regularization term。J=L+R先说损失项loss,再说regularization项。1. 分对得分1,分错得分0.gold standard2. hinge loss(for softmargin svm),J=1/2||w||^2 + sum(max(0,1-yf(w,x)))

2015-05-29 10:24:53 2097

原创 3.13 Anagrams

Notes: Given an array of strings, return all groups of strings that are anagrams. Note: All inputs will be in lower-case.  Solution: Sort the string to see if t

2015-05-28 18:03:27 417

原创 3.12CountandSay

Notes: The count-and-say sequence is the sequence of integers beginning as follows: 1, 11, 21, 1211, 111221, ...  1 is read off as "one 1" or 11. 11 i

2015-05-28 17:05:50 413

原创 3.9 ValidNumber

Notes: Validate if a given string is numeric. Some examples: "0" => true " 0.1 " => true "abc" => false "1 a" => false "2e10" => t

2015-05-27 17:45:42 368

原创 3.8LongestCommonPrefix

Notes: Write a function to find the longest common prefix string amongst an array of strings.  Solution: ... */class Solution {public: string longe

2015-05-27 17:09:59 369

原创 3.7WildcardMatching

Notes: Implement wildcard pattern matching with support for '?' and '*'. '?' Matches any single character. '*' Matches any sequence of characters (including the empty seq

2015-05-27 16:30:50 652

原创 3.6RegularExpressionMatching

Notes: Implement regular expression matching with support for '.' and '*'. '.' Matches any single character. '*' Matches zero or more of the preceding element. 

2015-05-27 14:52:53 448

原创 3.5LongestPalindromicSubstring

Notes: Given a string S, find the longest palindromic substring in S. You may assume that the maximum length of S is 1000, and there exists one unique longest palindromic substri

2015-05-27 13:29:06 267

原创 3.4 add binary

Notes: Given two binary strings, return their sum (also a binary string). For example, a = "11" b = "1" Return "100".  Solution: '

2015-05-26 17:12:38 411

原创 3.3StringtoInteger(atoi)

Notes: Implement atoi to convert a string to an integer. Hint: Carefully consider all possible input cases. If you want a challenge, please do not see below and ask you

2015-05-26 16:46:35 426

原创 3.2ImplementstrStr

Notes: Implement strStr(). Returns a pointer to the first occurrence of needle in haystack, or null if needle is not part of haystack.  Solution: 1. Check in th

2015-05-26 16:29:49 297

原创 3.1ValidPalindrome

Notes: Given a string, determine if it is a palindrome, considering only alphanumeric characters and ignoring cases. For example, "A man, a plan, a canal: Panama" is a pa

2015-05-26 16:14:09 365

原创 2.28SwapNodesinPairs

Notes: Given a linked list, swap every two adjacent nodes and return its head. For example, Given 1->2->3->4, you should return the list as 2->1->4->3. Your alg

2015-05-26 15:44:00 413

原创 2.2.10CopyListwithRandomPointer

Notes: A linked list is given such that each node contains an additional random pointer which could point to any node in the list or null. Return a deep copy of the lis

2015-05-26 15:37:13 274

原创 2.2.9ReverseNodesinkGroup

Notes: Given a linked list, reverse the nodes of a linked list k at a time and return its modified list. If the number of nodes is not a multiple of k then left-out nodes in the en

2015-05-25 17:51:56 412

原创 2.2.4RemoveDuplicatesfromSortedList

Notes: Given a sorted linked list, delete all duplicates such that each element appear only once. For example, Given 1->1->2, return 1->2. Given 1->1->2->3->3,

2015-05-25 16:49:54 306

原创 2.27RemoveNthNodeFromEndofList

Note: Given n will always be valid. Try to do this in one pass.  Solution: head---back------front------>NULL | | ---> n  */ 

2015-05-25 15:28:01 305

原创 2.2.6 RotateList.h

Notes: Given a list, rotate the list to the right by k places, where k is non-negative.  For example: Given 1->2->3->4->5->NULL and k = 2, return 4->5

2015-05-25 15:21:15 298

原创 2.2.5 RemoveDuplicatesfromSortedListII.h

Notes: Given a sorted linked list, delete all nodes that have duplicate numbers, leaving only distinct numbers from the original list. For example, Given 1->2->3->3->4->4

2015-05-25 15:07:47 395

原创 2.2.3 Partition List

Notes: Given a linked list and a value x, partition it such that all nodes less than x come before nodes greater than or equal to x. You should preserve the original relative order

2015-05-25 14:32:57 292

原创 2.2.2Reverse Linked ListII

Notes: Reverse a linked list from position m to n. Do it in-place and in one-pass. For example: Given 1->2->3->4->5->NULL, m = 2 and n = 4, return 1->4->3->2->5

2015-05-25 14:20:17 325

原创 对大型矩阵输出到终端和txt文档的方法

一般比较小的举证,直接用 print 命令就可以将举证输出到终端,对于比较大的矩阵,我的是500X500,我的方法如下:import numpy as npx = np.array([[1,2,4,9,10],[3,4,2,3,7]])print xw,h =x.shape#print w #print hfor i in range(0,w): word=x[i] f

2015-05-25 13:49:01 627

转载 CNN框架理解

9.5、Convolutional Neural Networks卷积神经网络       卷积神经网络是人工神经网络的一种,已成为当前语音分析和图像识别领域的研究热点。它的权值共享网络结构使之更类似于生物神经网络,降低了网络模型的复杂度,减少了权值的数量。该优点在网络的输入是多维图像时表现的更为明显,使图像可以直接作为网络的输入,避免了传统识别算法中复杂的特征提取和数据重建过程。卷积网络

2015-05-19 15:32:49 3254

转载 用GDB调试程序

原文:http://blog.csdn.net/haoel/article/details/2879GDB概述————GDB是GNU开源组织发布的一个强大的UNIX下的程序调试工具。或许,各位比较喜欢那种图形界面方式的,像VC、BCB等IDE的调试,但如果你是在UNIX平台下做软件,你会发现GDB这个调试工具有比VC、BCB的图形化调试器更强大的功能。所谓“寸有所长

2015-05-18 20:03:01 442

转载 《Notes on Convolutional Neural Networks》

转载与:http://blog.csdn.net/zouxy09/article/details/9993371一、介绍         这个文档讨论的是CNNs的推导和实现。CNN架构的连接比权值要多很多,这实际上就隐含着实现了某种形式的规则化。这种特别的网络假定了我们希望通过数据驱动的方式学习到一些滤波器,作为提取输入的特征的一种方法。         本文

2015-05-18 17:39:50 640

转载 Caffe导读之layer

转自http://www.shwley.com/index.php/archives/68/前言老实说,caffe中的layer层代码比较多,各种抽象看起来比较绕。官方关于Layer的教程写的很清楚,我根据这个文档,简单画了个图,再理解起来就方便了一些。layer.hpp和layer相关的头文件有:common_layers.hppdata_layers.hpplaye

2015-05-17 19:29:21 732

转载 caffe源码解析之blob

看过caffe官方文档的话,应该会知道,它可以分为三层:Blob、Layer、Net。Blob是一个四维的数组,用于存储数据,包括输入数据、输出数据、权值等等;Layer层则是神经网络中具体的各层结构,主要是计算的作用,在根据配置文件初始化结构后,前向计算结果,反向更新参数,都是它要做的,而它的输入和输出都是Blob数据;Net的话,就是多个Layer组合而成的有向无环图结构,也就是具体的网络了。

2015-05-17 19:20:54 594

转载 Caffe代码导读(1):Protobuf例子

原文章:http://blog.csdn.net/kkk584520/article/details/41046827Protobuf是一种可以实现内存与外存交换的协议接口。这是由谷歌开发的开源工具,目前研究Caffe源码时用到。一个软件项目 = 数据结构 + 算法 + 参数,对于数据结构和算法我们都已经有较多研究,但不同开发者对参数管理却各有千秋。有人喜欢TXT格式化

2015-05-17 17:19:23 440

原创 2.2.1Add Two Numbers

Notes:You are given two linked lists representing two non-negative numbers.The digits are stored in reverse order and each of their nodes contain a single digit.

2015-05-14 17:13:16 434

原创 2.1.20 Gas Station

Notes:There are N gas stations along a circular route, where the amount of gas at station i is gas[i].You have a car with an unlimited gas tank and it costs cost[i] of gas to t

2015-05-13 17:46:58 379

原创 2.1.19 Set Matrix Zeroes

Notes:Given a m x n matrix, if an element is 0, set its entire row and column to 0. Do it in place.Follow up:Did you use extra space?A straight forward

2015-05-13 17:33:22 498

原创 2.1.18Gray Code

Notes: The gray code is a binary numeral system where two successive values differ in only one bit. Given a non-negative integer n representing the total number of bits in the code

2015-05-12 17:25:35 452

原创 2.1.17 Climbing Stairs

Notes:You are climbing a stair case. It takes n steps to reach to the top.Each time you can either climb 1 or 2 steps. In how many distinct ways can you climb to the top?

2015-05-12 16:29:41 495

原创 2.1.16 Plus One

Notes: Given a number represented as an array of digits, plus one to the number.class Solution {public: vector plusOne(vector &digits) { int carry = 1, N = digits.size()

2015-05-12 16:10:50 455

原创 2.1.14 TrappingRainWater

notes:Given n non-negative integers representing an elevation map where the width ofeach bar is 1, compute how much water it is able to trap after raining.For exa

2015-05-11 17:17:32 481

原创 3Sum.h

/* Author: Annie Kim, anniekim.pku@gmail.com Date: Apr 19, 2013 Update: Sep 22, 2013 Problem: 3Sum Difficulty: Medium Source: http://leetcod

2015-05-07 16:26:58 614

空空如也

空空如也

TA创建的收藏夹 TA关注的收藏夹

TA关注的人

提示
确定要删除当前文章?
取消 删除