<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[inner_sum的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/inner_sum</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; inner_sum]]></copyright><item><title><![CDATA[faster-rcnn用于训练人脸识别数据集wider_face]]></title><link>https://blog.youkuaiyun.com/inner_sum/article/details/80628986</link><guid>https://blog.youkuaiyun.com/inner_sum/article/details/80628986</guid><author>inner_sum</author><pubDate>Fri, 08 Jun 2018 22:56:20 +0800</pubDate><description><![CDATA[数据接口修改：参考博客（caffe版本与tf版本工程结构一致，可相互借鉴）：http://jacobkong.github.io/posts/464905881/http://jacobkong.github.io/posts/4113466123/用于自己标注的数据时出现额外报错如下：Nan in summary histogram xxx：错误来源：数据Annotations中xml文件标记框...]]></description><category></category></item><item><title><![CDATA[交叉熵以及为什么用交叉熵作为代价函数]]></title><link>https://blog.youkuaiyun.com/inner_sum/article/details/79511023</link><guid>https://blog.youkuaiyun.com/inner_sum/article/details/79511023</guid><author>inner_sum</author><pubDate>Sat, 10 Mar 2018 19:58:23 +0800</pubDate><description><![CDATA[作者：知乎用户链接：https://www.zhihu.com/question/41252833/answer/108777563来源：知乎著作权归作者所有。商业转载请联系作者获得授权，非商业转载请注明出处。熵的本质是香农信息量()的期望。现有关于样本集的2个概率分布p和q，其中p为真实分布，q非真实分布。按照真实分布p来衡量识别一个样本的所需要的编码长度的期望(即平均编码长度)为：H(p)=。...]]></description><category></category></item><item><title><![CDATA[jsp中显示本地图片]]></title><link>https://blog.youkuaiyun.com/inner_sum/article/details/50299383</link><guid>https://blog.youkuaiyun.com/inner_sum/article/details/50299383</guid><author>inner_sum</author><pubDate>Mon, 14 Dec 2015 19:20:00 +0800</pubDate><description><![CDATA[jsp显示本地文件]]></description><category></category></item></channel></rss>