<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[langxing1992的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/langxing1992</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; langxing1992]]></copyright><item><title><![CDATA[ubuntu下tensorflow的环境搭建]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/77099652</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/77099652</guid><author>langxing1992</author><pubDate>Fri, 11 Aug 2017 16:28:09 +0800</pubDate><description><![CDATA[ubuntu下tensorflow的环境搭建


转自：http://blog.youkuaiyun.com/cxq234843654/article/details/70857562
https://www.leiphone.com/news/201606/ORlQ7uK3TIW8xVGF.html



官网参考文档地址：https://www.tensorflow.org/install]]></description><category></category></item><item><title><![CDATA[Caffe 特征图可视化]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/74905578</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/74905578</guid><author>langxing1992</author><pubDate>Sun, 09 Jul 2017 21:41:32 +0800</pubDate><description><![CDATA[特征图即网络在前向传播阶段的各层输出。 http://blog.youkuaiyun.com/u012938704/article/details/52767695
这里使用一个训练过的CaffeNet模型对一张猫的图片进行分类，使用matlab绘制出卷积层的特征图。
今天在使用matcaffe时遇到了CUDNN_STATUS_ARCH_MISMATCH的错误，根据http://blog.youkuaiyun.com]]></description><category></category></item><item><title><![CDATA[yolo]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72784251</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72784251</guid><author>langxing1992</author><pubDate>Sat, 27 May 2017 20:37:00 +0800</pubDate><description><![CDATA[【YOLO学习】使用YOLO v2训练自己的数据
作者：hysteric314

说明
这篇文章是训练YOLO v2过程中的经验总结，我使用YOLO v2训练一组自己的数据，训练后的model，在阈值为.25的情况下，Recall值是95.54%，Precision 是97.27%。

需要注意的是，这一训练过程可能只对我自己的训练集有效，因为我是根据我这一训练集的特征来对YOLO代码]]></description><category></category></item><item><title><![CDATA[spyder绘制loss曲线]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72764130</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72764130</guid><author>langxing1992</author><pubDate>Fri, 26 May 2017 10:30:26 +0800</pubDate><description><![CDATA[caffe的python接口学习（7）：绘制loss和accuracy曲线
转自：http://www.cnblogs.com/denny402/p/5686067.html

使用python接口来运行caffe程序，主要的原因是python非常容易可视化。所以不推荐大家在命令行下面运行python程序。如果非要在命令行下面运行，还不如直接用 c++算了。
推荐使用jupyter not]]></description><category></category></item><item><title><![CDATA[python caffe报错：No module named google.protobuf.internal]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72596920</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72596920</guid><author>langxing1992</author><pubDate>Sun, 21 May 2017 11:44:54 +0800</pubDate><description><![CDATA[http://blog.youkuaiyun.com/wuzuyu365/article/details/52431062


Python caffe报错：No module named google.protobuf.internal
我装的是anaconda2, 解决方法是在其中安装protobuf最新版本
sudo chmod 777 -R anaconda2
conda install ]]></description><category></category></item><item><title><![CDATA[ssd]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72594133</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72594133</guid><author>langxing1992</author><pubDate>Sun, 21 May 2017 09:59:41 +0800</pubDate><description><![CDATA[摘要: 配置ssd 并且整理自己的数据进行训练，调用c++ 接口来做检测分类 
最近一直在看深度学习的东西，用到了ssd，上网一搜，找到一篇不错的博客，http://www.cnblogs.com/objectDetect/p/5780006.html，官网，这篇是连安装cuda到caffe环境的一系列http://blog.youkuaiyun.com/wopawn/article/details/523]]></description><category></category></item><item><title><![CDATA[解决faster-rcnn中训练时assert（boxes[:,2]>=boxes[:,0]）.all()的问题]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72568867</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72568867</guid><author>langxing1992</author><pubDate>Sat, 20 May 2017 09:41:32 +0800</pubDate><description><![CDATA[1、出现问题：训练faster rcnn时出现如下报错：
File "/py-faster-rcnn/tools/../lib/datasets/imdb.py", line 108, in append_flipped_images
    assert (boxes[:, 2] >= boxes[:, 0]).all()
AssertionError
2、问题分析：
检查自己数据发现，左]]></description><category></category></item><item><title><![CDATA[fasterrcnn训练自己的数据]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72566350</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72566350</guid><author>langxing1992</author><pubDate>Fri, 19 May 2017 21:57:32 +0800</pubDate><description><![CDATA[http://blog.youkuaiyun.com/sinat_30071459/article/details/51332084


说明：本博文假设你已经做好了自己的数据集，该数据集格式和VOC2007相同。下面是训练前的一些修改。
（做数据集的过程可以看http://blog.youkuaiyun.com/sinat_30071459/article/details/50723212）


Fast]]></description><category></category></item><item><title><![CDATA[将数据集做成VOC2007格式用于Faster-RCNN训练]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72566335</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72566335</guid><author>langxing1992</author><pubDate>Fri, 19 May 2017 21:55:31 +0800</pubDate><description><![CDATA[转自：http://blog.youkuaiyun.com/sinat_30071459/article/details/50723212

0.文件夹名
首先，确定你的数据集所放的文件夹名字，例如我的叫logos。
（或者和voc2007一样的名字：VOC2007）
1.图片命名
虽然说图片名对训练没什么影响，但建议还是按VOC2007那样，如“000005.jpg”这种形式。至于图片格式，代码]]></description><category></category></item><item><title><![CDATA[fast-rcnn训练自己的数据]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72566322</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72566322</guid><author>langxing1992</author><pubDate>Fri, 19 May 2017 21:53:22 +0800</pubDate><description><![CDATA[转自：.https://github.com/coldmanck/fast-rcnn

(Concerning to original README.md, please refer to README_ORIG.md.)
Training Fast R-CNN on
Right Whale Recognition Dataset
Before starting to train you]]></description><category></category></item><item><title><![CDATA[caffe绘制训练过程的loss和accuracy曲线]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72453127</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72453127</guid><author>langxing1992</author><pubDate>Wed, 17 May 2017 21:25:50 +0800</pubDate><description><![CDATA[转自：http://blog.youkuaiyun.com/u013078356/article/details/51154847
在caffe的训练过程中，大家难免想图形化自己的训练数据，以便更好的展示结果。如 果自己写代码记录训练过程的数据，那就太麻烦了，caffe中其实已经自带了这样的小工具 caffe-master/tools/extra/parse_log.sh  caffe-master/too]]></description><category></category></item><item><title><![CDATA[yolo训练自己的数据]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/72084315</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/72084315</guid><author>langxing1992</author><pubDate>Sun, 14 May 2017 23:11:07 +0800</pubDate><description><![CDATA[转载：   http://blog.youkuaiyun.com/ch_liu23/article/details/53558549

最近在用yolo来做视频中的人员检测，选择YOLO是从速度考虑，当然也可以用ssd。YOLO相关可看主页Darknet，有相关代码和使用方法。由于之前做自己的数据训练过程中出现各种问题，参照了各种博客才跑通，现在记录下以防后面忘记，也方便自己总结一下。
      YO]]></description><category></category></item><item><title><![CDATA[fast-rcnn中编译cudnn出错解决方法]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/71440831</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/71440831</guid><author>langxing1992</author><pubDate>Tue, 09 May 2017 11:59:11 +0800</pubDate><description><![CDATA[faster rcnn 代码默认是使用的cudnn v4, 但是为了体验最新的v5, 或者使用GTX1080 ,我们编译faster rcnn的时候就会报错：

In file included from ./include/caffe/util/cudnn.hpp:5:0,
                 from ./include/caffe/util/device_alternat]]></description><category></category></item><item><title><![CDATA[如何运行python文件]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/71367833</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/71367833</guid><author>langxing1992</author><pubDate>Sun, 07 May 2017 20:22:10 +0800</pubDate><description><![CDATA[先将终端所在路径切换到python脚本文件的目录下
     然后给脚本文件运行权限，一般755就OK，如果完全是自己的私人电脑，也不做服务器什么的，给777的权限问题也不大(具体权限含义参考chmod指令的介绍，就不赘述了)：
    chmod 755 ./*.py
    然后执行。
    如果在脚本内容的开头已经给出了类似于如下的注释：
    #!/usr/bin/env python]]></description><category></category></item><item><title><![CDATA[测试python代码]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/71367738</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/71367738</guid><author>langxing1992</author><pubDate>Sun, 07 May 2017 20:19:22 +0800</pubDate><description><![CDATA[#!/usr/bin/env python
#coding=utf-8  
#加载必要的库  
import numpy as np  
import sys,os  
#设置当前目录  
caffe_root = '/home/cjj/caffe/'    
sys.path.insert(0, caffe_root + 'python')  
import caffe  
o]]></description><category></category></item><item><title><![CDATA[生成train.txt和val.txt]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/71279539</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/71279539</guid><author>langxing1992</author><pubDate>Sat, 06 May 2017 19:12:58 +0800</pubDate><description><![CDATA[#!/usr/bin/env sh
DATA=/home/cjj/caffe/data/myimage
MY=/home/cjj/caffe/data/myimage
echo "Create train.txt..."
rm -rf $MY/train.txt
for i in 3 4 5 6 7 
do
find $DATA/train -name $i*.jpg | cut -]]></description><category></category></item><item><title><![CDATA[自动生成图像标签代码]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/71279326</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/71279326</guid><author>langxing1992</author><pubDate>Sat, 06 May 2017 19:09:36 +0800</pubDate><description><![CDATA[%%  
    %图片保存路径为：  
    %E:\image\car  
    %E:\image\person  
    %car和person是保存车和行人的文件夹  
    %这些文件夹还可以有多个，  
    %放在image文件夹里就行  
    %该代码的作用是将图片名字改成000123.jpg这种形式  
    %%  
    %clc;  ]]></description><category></category></item><item><title><![CDATA[linux caffe环境配置]]></title><link>https://blog.youkuaiyun.com/langxing1992/article/details/70807856</link><guid>https://blog.youkuaiyun.com/langxing1992/article/details/70807856</guid><author>langxing1992</author><pubDate>Wed, 26 Apr 2017 16:30:37 +0800</pubDate><description><![CDATA[1.环境准备
sudo apt-get install git





sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev
 libhdf5-serial-dev protobuf-compiler 

sudo apt-get install --no-install]]></description><category></category></item></channel></rss>