<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[LBJ_python的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/LBJ_python</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; LBJ_python]]></copyright><item><title><![CDATA[安装 虚拟 环境]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/90268842</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/90268842</guid><author>LBJ_python</author><pubDate>Thu, 16 May 2019 17:55:58 +0800</pubDate><description><![CDATA[安装虚拟环境
sudo pip install virtualenv
创建文件
virtualenv -p /usr/bin/pythonX xxxx-env
开启进入虚拟环境
source ~/xxxx-env/bin/activate
退出
deactivate

]]></description><category></category></item><item><title><![CDATA[Ubuntu 系统换源 pip 换源]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/90268684</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/90268684</guid><author>LBJ_python</author><pubDate>Thu, 16 May 2019 17:49:54 +0800</pubDate><description><![CDATA[更换Ubuntu的系统源和Python的pip源能极大的提高下载速度，是安装系统后必不可少的一项工作。一、更换系统源1、复制原文件备份sudo cp /etc/apt/sources.list /etc/apt/sources.list.bak2、编辑源列表文件s...]]></description><category></category></item><item><title><![CDATA[esc 4000 G2 磁盘阵列配置]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/86536771</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/86536771</guid><author>LBJ_python</author><pubDate>Fri, 18 Jan 2019 11:16:36 +0800</pubDate><description><![CDATA[根据技术师傅指导,自己配置esc 4000 G2 磁盘阵列 拍图片进行保存,以便以后使用,还是对其不是很明白
贴图记录下,开启服务器,服务器正常检测到阵列卡,ctrl+h 进入webbios界面,右侧 logical view 会显示检测到的磁盘信息
图片顺序即是配置顺序:
到这里就配置完成了,还有添加和删除磁盘,应该都是类似的,还有尝试!

...]]></description><category></category></item><item><title><![CDATA[Batch Normalization 中use_global_stats参数]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/86063914</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/86063914</guid><author>LBJ_python</author><pubDate>Tue, 08 Jan 2019 10:52:06 +0800</pubDate><description><![CDATA[Batch Normalization层中有个参数use_global_stats，它表示是否使用caffe内部的均值和方差。训练的时候，要将use_global_stats设置为false，而测试的时候将use_global_stats设置为true
否则:训练的时候  loss值会一直保持在87.3365,而且不会变

...]]></description><category></category></item><item><title><![CDATA[【Linux】下分解大压缩文件或解压拆分后的文件]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/85779615</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/85779615</guid><author>LBJ_python</author><pubDate>Fri, 04 Jan 2019 15:39:42 +0800</pubDate><description><![CDATA[unzip testall.zip`# 1.先合并文件
cat test.z* &amp;gt; testall.zip # 2.解压


有没有遇到某些网站上传复件时，单个文件有大小限制，导致上传失败呢？当然你可以采用更高的压缩率重新生成压缩包来解决，但如果还是超出大小限制，那该怎么办呢？有个方法可以解决此类问题，将压缩包分成多个文件，问题就迎刃而解，这就叫：分卷压缩。
如何进行分卷压缩与解压呢，有如下...]]></description><category></category></item><item><title><![CDATA[caffe 中 BatchNorm layer设定]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/84853262</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/84853262</guid><author>LBJ_python</author><pubDate>Thu, 06 Dec 2018 14:38:54 +0800</pubDate><description><![CDATA[
转载于http://blog.youkuaiyun.com/u012939857/article/details/70740283

BN层的设定一般是按照conv-&amp;gt;bn-&amp;gt;scale-&amp;gt;relu的顺序来形成一个block。&amp;nbsp;
关于bn，有一个注意点，caffe实现中的use_global_stats参数在训练时设置为false，测试时设置为true。&amp;nbsp;
因为在训练时...]]></description><category></category></item><item><title><![CDATA[python中OrderedDict的使用]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/84248409</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/84248409</guid><author>LBJ_python</author><pubDate>Mon, 19 Nov 2018 14:34:29 +0800</pubDate><description><![CDATA[很多人认为python中的字典是无序的，因为它是按照hash来存储的，但是python中有个模块collections(英文，收集、集合)，里面自带了一个子类
OrderedDict，实现了对字典对象中元素的排序。请看下面的实例
import collections
print &quot;Regular dictionary&quot;
d={}
d['a']='A'
d['b']='B'
d['c']='C'
f...]]></description><category></category></item><item><title><![CDATA[python中defaultdict的用法详解]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/84248107</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/84248107</guid><author>LBJ_python</author><pubDate>Mon, 19 Nov 2018 14:29:55 +0800</pubDate><description><![CDATA[
							
							
							认识defaultdict：
当我使用普通的字典时，用法一般是dict={},添加元素的只需要dict[element] =value即，调用的时候也是如此，dict[element] = xxx,但前提是element字典里，如果不在字典里就会报错，如： 
 
这时defaultdict就能排上用场了，defaultdict的作用是在于，当字典里的...]]></description><category></category></item><item><title><![CDATA[SSD: Single Shot MultiBox Detector 模型fine-tune和网络架构]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83867349</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83867349</guid><author>LBJ_python</author><pubDate>Thu, 08 Nov 2018 20:07:07 +0800</pubDate><description><![CDATA[
							
							
							前言 
博主在上一篇中提到了两种可能的改进方法。其中方法1，扩充类似数据集，详见Udacity Self-Driving 目标检测数据集简介与使用 ，由于一些原因，并未对此数据集做过多探索，一次简单训练下，mAP为64%左右，这还需要进一步探索。而方法2，说的是fine-tune已经训练好的SSD model，最近没来得及进行调参，初次实验效果有限，...]]></description><category></category></item><item><title><![CDATA[linux 下查看机器是cpu是几核的]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83793650</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83793650</guid><author>LBJ_python</author><pubDate>Tue, 06 Nov 2018 20:24:48 +0800</pubDate><description><![CDATA[linux 下查看机器是cpu是几核的
几个cpu
more /proc/cpuinfo |grep “physical id”|uniq|wc -l
每个cpu是几核（假设cpu配置相同）
more /proc/cpuinfo |grep “physical id”|grep “0”|wc -l
cat /proc/cpuinfo | grep processor


查看物理CPU的个数
#c...]]></description><category></category></item><item><title><![CDATA[caffe 安装问题小总结（make pycaffe]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83783278</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83783278</guid><author>LBJ_python</author><pubDate>Tue, 06 Nov 2018 10:55:12 +0800</pubDate><description><![CDATA[caffe 安装问题小总结（make pycaffe）
1\ Make pycaffe -j8编译报错
gedit ~/.bashrc打开文件,最后面加入 自己的路径
export PYTHONPATH=/home/hyhuang/下载/caffe/python
source ~/.bashrc,生成修改
错误修改1,caffe目录下,进入Python环境
python/caffe/_caffe....]]></description><category></category></item><item><title><![CDATA[P-R曲线与mAP编辑器]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83684758</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83684758</guid><author>LBJ_python</author><pubDate>Sat, 03 Nov 2018 10:30:32 +0800</pubDate><description><![CDATA[
								            
						

reference：http://blog.youkuaiyun.com/marising/article/details/6543943
在信息检索、分类体系中，有一系列的指标，搞清楚这些指标对于评价检索和分类性能非常重要，因此最近根据网友的博客做了一个汇总。

准确率、召回率、F1

信息检索、分类、识别、翻译等领域两个最基本指标是召回率(Rec...]]></description><category></category></item><item><title><![CDATA[mysql 小demo]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83652014</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83652014</guid><author>LBJ_python</author><pubDate>Fri, 02 Nov 2018 10:55:23 +0800</pubDate><description><![CDATA[
#! -*- coding:utf-8 -*-

import pymysql
import os
import time

class Mysql(object):
    def __init__(self,host='localhost',port=3306,user='root',password='root123',db='hongxi'):
        self.host = h...]]></description><category></category></item><item><title><![CDATA[cifar10 数据集转换成图片]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83651739</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83651739</guid><author>LBJ_python</author><pubDate>Fri, 02 Nov 2018 10:47:32 +0800</pubDate><description><![CDATA[自己亲测，转换一部分，可以使用


import cv2
import numpy as np


file = 'data_batch_2'

官方给的示例：


def unpickle(file):
    import cPickle
    with open(file, 'rb') as fo:
        dict = cPickle.load(fo)
    return di...]]></description><category></category></item><item><title><![CDATA[安装caffe   /usr/bin/ld: cannot find -lhdf5]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83626333</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83626333</guid><author>LBJ_python</author><pubDate>Thu, 01 Nov 2018 20:27:01 +0800</pubDate><description><![CDATA[ 每次安装都是新的体验：

如果编译caffe时提示找不到libhdf5.so , 错误提示为：/usr/bin/ld: cannot find -lhdf5
使用命令sudo find / -name libhdf5.so ， 我找到的路径为：
/usr/lib/x86_64-linux-gnu/hdf5/serial/libhdf5.so , 此时要修改Makefile.config中的项“L...]]></description><category></category></item><item><title><![CDATA[基于自制数据集的MobileNet-SSD模型训练]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83626265</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83626265</guid><author>LBJ_python</author><pubDate>Thu, 01 Nov 2018 20:18:10 +0800</pubDate><description><![CDATA[基于自制数据集的MobileNet-SSD模型训练

来源：QQ快报  责任编辑:小易  

“本文主要内容：基于自制的仿VOC数据集，利用caffe框架下的MobileNet-SSD模型训练。”

本文的base是https://github.com/chuanqi305/MobileNet-SSD，这个project又是基于https://github.com/weiliu89/caffe/t...]]></description><category></category></item><item><title><![CDATA[cifar10训练 test 0.894]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83626203</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83626203</guid><author>LBJ_python</author><pubDate>Thu, 01 Nov 2018 20:15:38 +0800</pubDate><description><![CDATA[name: &quot;CIFAR10_full&quot;
layer {
  name: &quot;cifar&quot;
  type: &quot;Data&quot;
  top: &quot;data&quot;
  top: &quot;label&quot;
  include {
    phase: TRAIN
  }
  transform_param {
    mean_file: &quot;mean.binaryproto&quot;
    #scale:0.0039062]]></description><category></category></item><item><title><![CDATA[mysql]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83536950</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83536950</guid><author>LBJ_python</author><pubDate>Tue, 30 Oct 2018 09:38:50 +0800</pubDate><description><![CDATA[

 






insert into t2(id,name,age) values(%s,%s,%s) on duplicate key update name=values(name),age=values(age)

已有表  t2 ,字段类型：id  int(2) ;  name char(10) ; age int(2) id 为 aotu_increment primary key...]]></description><category></category></item><item><title><![CDATA[caffe 前向传播和反向传播]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83444572</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83444572</guid><author>LBJ_python</author><pubDate>Sat, 27 Oct 2018 11:18:51 +0800</pubDate><description><![CDATA[caffe学习笔记3从3.1开始主要翻译一下caffe的官方文档，写的非常好，忍不住要作一下。本篇翻译来自caffe官网的：http://caffe.berkeleyvision.org/tutorial/forward_backward.html

前向传播和反向传播是计算神经网络非常重要的部分。



考虑一个简单的逻辑回归分类问题

前向传播：通过网络的输入计算输出结果的过程，在前向传播中，...]]></description><category></category></item><item><title><![CDATA[caffe+报错︱深度学习参数调优杂记+caffe训练时的问题+dropout/batch Normalization]]></title><link>https://blog.youkuaiyun.com/LBJ_python/article/details/83268300</link><guid>https://blog.youkuaiyun.com/LBJ_python/article/details/83268300</guid><author>LBJ_python</author><pubDate>Mon, 22 Oct 2018 09:24:19 +0800</pubDate><description><![CDATA[https://blog.youkuaiyun.com/sinat_26917383/article/details/54232791]]></description><category></category></item></channel></rss>