<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
	<channel>
		<atom:link href="http://itecfun.com/extern.php?action=feed&amp;fid=59&amp;type=rss" rel="self" type="application/rss+xml" />
		<title><![CDATA[个人知识管理站 / 机器学习]]></title>
		<link>http://www.itecfun.com/index.php</link>
		<description><![CDATA[个人知识管理站 最近发表的主题。]]></description>
		<lastBuildDate>Tue, 27 Aug 2019 00:49:59 +0000</lastBuildDate>
		<generator>FluxBB</generator>
		<item>
			<title><![CDATA[IT 行业一些常用术语]]></title>
			<link>http://www.itecfun.com/viewtopic.php?id=3412&amp;action=new</link>
			<description><![CDATA[<p>1. <strong>Scalability</strong> – The architecture of your cloud computing system can be easily enlarged to accommodate<br />the growth of users in the future. That means when your application attracts much more users, you<br />need to spend more money on infrastructures like computers and network bandwidth, nearly no cost<br />on development.<br />2.<strong> High Availability </strong>– The architecture of your cloud computing system can provide reliable and<br />consistent services. The architecture of your system can guarantee 24/7/365 and 99.99% availability.<br />High availability can, from a massive pool of redundant IT resources, as well as quick failover<br />mechanism - if a server fails, hosted applications and services can easily be transited to any of the<br />available servers.<br />3. <strong>Low Maintenance </strong>– The architecture of your cloud computing system provides enhanced and<br />simplified IT management and maintenance capabilities through central administration of resources.<br />IT infrastructure updates and maintenance are eliminated, as all resources are maintained by cloud<br />computing service providers.</p>]]></description>
			<author><![CDATA[dummy@example.com (xuyg)]]></author>
			<pubDate>Tue, 27 Aug 2019 00:49:59 +0000</pubDate>
			<guid>http://www.itecfun.com/viewtopic.php?id=3412&amp;action=new</guid>
		</item>
		<item>
			<title><![CDATA[python3.6  + tesseract-orc 4.0 + pillow 5.1图片识别（中文识别）(源码 可执行）【原创】]]></title>
			<link>http://www.itecfun.com/viewtopic.php?id=3375&amp;action=new</link>
			<description><![CDATA[<p>tesseract4.0安装参考<a href="https://blog.csdn.net/a745233700/article/details/80108652" rel="nofollow">链接进入</a></p><p>以下代码保存文件名为trans2Chi_sim.py</p><div class="codebox"><pre class="vscroll"><code>#!/usr/bin/env python
# -*- coding: utf-8 -*-

from PIL import Image
import pytesseract

import argparse

#命令行执行 添加图片目录和结果保存路径
ap = argparse.ArgumentParser()

ap.add_argument(&quot;-i&quot;,&quot;--image&quot;,required=True,help=&quot;path to input image&quot;)
#结果保存的文件可以不设置，默认保存到当前目录下的result.txt文件中
ap.add_argument(&quot;-r&quot;,&quot;--result&quot;,required=False,help=&quot;the name of the result txt file save name&quot;)

args = vars(ap.parse_args())

#获取图片
img_path = args[&quot;image&quot;]
print(&#039;image path : %s&#039;%img_path)

#识别结果默认保存到当前目录result.txt文件中
res_path =&#039;result&#039;
if args[&quot;result&quot;] is not None:
	res_path = args[&quot;result&quot;]

#print(&#039;result path : %s&#039;%res_path)

image = Image.open(img_path)
#使用tessdata chi_sim简体中文训练数据进行识别
text = pytesseract.image_to_string(image, lang=&#039;chi_sim&#039;)

filename = res_path
if &#039;.txt&#039; not in res_path:
	filename = res_path+&quot;.txt&quot;
#结果保存
with open(filename,&#039;w&#039;) as f:
	f.write(text.replace(&#039; &#039;,&#039;&#039;))  #识别后每个字之间有空格，需要去掉空格

print(&#039;识别内容已保存到文件：%s&#039;%filename)</code></pre></div><br /><p>命令号执行如下：<br />python trans2Chi_sim.py -i imagename.jpg [ -r result.txt]</p><p>非手写字体识别率非常高，很不错</p>]]></description>
			<author><![CDATA[dummy@example.com (xuyg)]]></author>
			<pubDate>Mon, 26 Nov 2018 02:50:59 +0000</pubDate>
			<guid>http://www.itecfun.com/viewtopic.php?id=3375&amp;action=new</guid>
		</item>
		<item>
			<title><![CDATA[Tensorflow 问题及解决方案]]></title>
			<link>http://www.itecfun.com/viewtopic.php?id=3372&amp;action=new</link>
			<description><![CDATA[<p>环境：windows10(64bit)+python3.6 + tensorflow 1.8.0 </p><p>tensorflow 通过pip install tensorflow 安装完成</p><p>1、Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA<br />问题：你的CPU支持AVX扩展，但你安装的TensorFlow版本无法编译使用。<br />原因是当前TensorFlow的版本不支持cpu的AVX2编译<br />可能是因为安装时使用的pip install tensorflow ，这样默认会下载X86_64的SIMD版本。<br />两种解决办法：<br />&#160; &#160;1）.忽略这个警告，不看它！<br />&#160; <strong> import os<br />&#160; &#160;os.environ[&quot;TF_CPP_MIN_LOG_LEVEL&quot;]=&#039;1&#039; # 这是默认的显示等级，显示所有信息&#160; <br />&#160; &#160;os.environ[&quot;TF_CPP_MIN_LOG_LEVEL&quot;]=&#039;2&#039; # 只显示 warning 和 Error&#160; &#160;<br />&#160; &#160;os.environ[&quot;TF_CPP_MIN_LOG_LEVEL&quot;]=&#039;3&#039; # 只显示 Error</strong></p><p>2.彻底解决，换成支持cpu用AVX2编译的TensorFlow版本。<br />首先，卸载原来版本的TensorFlow<br /><strong><br />pip uninstall tensorflow<br />pip uninstall tensorboard<br />pip uninstall protobuf<br />pip uninstall keras<br /></strong><br />然后到D:\Program Files\Python\Python36\Lib\site-packages 目录下 删除对应的文件夹 tensorflow keras 等</p><p>完成后可以用pip list查看一下所有库，检查是否成功卸载。</p><p>然后去github下载正确的tf版本，Windows<a href="https://github.com/fo40225/tensorflow-windows-wheel" rel="nofollow">点这里下载</a>。其他操作系统可以点这里找到对应的版本。</p><p>这里下载1.11.0\py36\CPU\avx2 版本（tensorflow-1.11.0-cp36-cp36m-win_amd64.whl）</p><br /><p>下载后安装即可：<br /><strong><br />pip install tensorflow-1.11.0-cp36-cp36m-win_amd64.whl </strong></p><p>安装完成再运行程序，执行成功不会报错</p>]]></description>
			<author><![CDATA[dummy@example.com (xuyg)]]></author>
			<pubDate>Wed, 31 Oct 2018 03:50:10 +0000</pubDate>
			<guid>http://www.itecfun.com/viewtopic.php?id=3372&amp;action=new</guid>
		</item>
	</channel>
</rss>
