<?xml version="1.0" encoding="utf-8" ?><rss version="2.0"><channel><title><![CDATA[AI_BN的博客]]></title><description><![CDATA[]]></description><link>https://blog.youkuaiyun.com/AI_BN</link><language>zh-cn</language><generator>https://blog.youkuaiyun.com/</generator><copyright><![CDATA[Copyright &copy; AI_BN]]></copyright><item><title><![CDATA[跟着仙凡兄学习编译Telegram vs2022 2024.1.11编译成功]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/135539926</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/135539926</guid><author>AI_BN</author><pubDate>Thu, 11 Jan 2024 22:04:59 +0800</pubDate><description><![CDATA[3.2 运行命令：cd tdesktop\Telegram\build\prepare。打开vpx.sln，修改如下，就可以编译了，分别编译debug和release。方法：需要将libvpx拷倒tg_owt文件夹，如果有问题，可以咨询。本人花了两天，问官方作者终于编译成功Telegram。也是由于第一个问题造成的，编译缺少libvpx静态库。如下：API_ID和API_HASH注册地址。策略：可以生成libvpx.sln。3.3 运行命令：win.bat。5 运行Telegram.sln。]]></description><category></category></item><item><title><![CDATA[ImportError: DLL load failed while importing imaging: 找不到指定的模块]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/135382368</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/135382368</guid><author>AI_BN</author><pubDate>Thu, 04 Jan 2024 11:22:13 +0800</pubDate><description><![CDATA[参考 https://blog.youkuaiyun.com/qq_45510888/article/details/121446878。参考：https://zhuanlan.zhihu.com/p/662305030?官网查询 https://pypi.org/project/pillow/#files。4）： 如果重装pillow的话，还是报错，降低pillow版本。说明：Pillow 与pillow都可以，不区分首字母大小写。第一步：查看是否有pillow。2）查看pillow的版本。3）：重装pillow。]]></description><category></category></item><item><title><![CDATA[跟着仙凡兄学习 ros2 ubuntu与windows的安装（本人小白开始，一步一步跟着坑走，最后都成功的）]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/135382273</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/135382273</guid><author>AI_BN</author><pubDate>Thu, 04 Jan 2024 11:19:13 +0800</pubDate><description><![CDATA[目  录1 ROS安装的步骤	21.1 ROS UBUNTU安装	21.1.1 环境配置：UBUNTU20.04	21.1.2 设置编码	21.1.3 添加ROS2 APT仓库	31.1.4 接下来可以安装ROS2了	31.1.5 创建 COLCON 工作区并下载教程	31.1.6 设置环境	41.1.7 双环境设置	41.1.8 验证之一	51.2 ROS2 WINDOWS安装	51.2.1 进入如下网站，安装CHOCOLATEY	5。]]></description><category></category></item><item><title><![CDATA[Windows远程桌面连线显示请稍后]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/126262085</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/126262085</guid><author>AI_BN</author><pubDate>Wed, 10 Aug 2022 10:59:53 +0800</pubDate><description><![CDATA[帮助远程桌面连接]]></description><category></category></item><item><title><![CDATA[windows vs2019libtorch不能成功调用cuda]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/122126146</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/122126146</guid><author>AI_BN</author><pubDate>Fri, 24 Dec 2021 13:35:00 +0800</pubDate><description><![CDATA[需要新加命令行/INCLUDE:?warp_size@cuda@at@@YAHXZ，目前版本 libtorch-win-shared-with-deps-1.6.0+cu101，亲测可用。

]]></description><category></category></item><item><title><![CDATA[部署onnx]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/122125884</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/122125884</guid><author>AI_BN</author><pubDate>Fri, 24 Dec 2021 13:17:24 +0800</pubDate><description><![CDATA[部署框架onnx
1.转成C++程序，输出要考虑清楚
2.cuda的库要匹配，主要还是与自己电脑上的驱动有关系，如果程序卡在cuda加载的时候，优先考虑是否cuda_adv_infer64_8.dll是否加载正确。
3 输入要正确 训练与推理要保持一致

...]]></description><category></category></item><item><title><![CDATA[c++程序未成功启动]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/122121903</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/122121903</guid><author>AI_BN</author><pubDate>Fri, 24 Dec 2021 10:07:55 +0800</pubDate><description><![CDATA[1 动态库缺少，比较明显；
2 动态库版本有问题
问题1：使用三方库
1）加载联调过程中的三方库，凡是修改第三方库关于头文件的库，都得重新生成一份程序
2）库冲突，比如不同的cuda库之间的冲突
问题二：版本不匹配
有时候程序莫名其妙，有可能第三方库的版本有问题，比如调用cv450可以，调用cv453就不可以

...]]></description><category></category></item><item><title><![CDATA[C++动态库使用STL，出现内存错误]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/121543233</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/121543233</guid><author>AI_BN</author><pubDate>Thu, 25 Nov 2021 17:43:28 +0800</pubDate><description><![CDATA[1 debug版本
设置运行库的方法

2  主程序传递参数给dll
dll中的参数都需要用引用
2 release
设置运行库的方法


]]></description><category></category></item><item><title><![CDATA[linux安装cuda]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/120117896</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/120117896</guid><author>AI_BN</author><pubDate>Sun, 05 Sep 2021 17:16:59 +0800</pubDate><description><![CDATA[1 使用nvidia-smi 和nvcc -V命令检查是否安装驱动以及cuda
2 如果没有直接去nvidia官网下载与自己电脑显卡匹配的驱动
cuda同样道理

]]></description><category></category></item><item><title><![CDATA[win10安装dcnv2]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/119874810</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/119874810</guid><author>AI_BN</author><pubDate>Mon, 23 Aug 2021 18:07:06 +0800</pubDate><description><![CDATA[亲自安装
1 git download dcnv2
网址：下载
2 打开vs215

3 cd 到dcn2下载的路径
在窗口运行python setup build develop
注意：
torch 1.8的环境失败，本人试的是torch1.3

]]></description><category></category></item><item><title><![CDATA[pem密钥（window，linux)]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/115699310</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/115699310</guid><author>AI_BN</author><pubDate>Wed, 14 Apr 2021 15:09:05 +0800</pubDate><description><![CDATA[1：本地传送到服务器中
scp -i my.pem xxxx root@47.104.139.50:/root
my.pem 表示阿里云的密钥
xxxx 本地文件
47.104.139.50 阿里云ip
/root 服务器文件
2：服务器传送到本地
scp -i my_key.pem root@201.200.167.22:/root/file1.txt /root/Desktop/file1.txt
window:
可以使用git bash直接用 xshell没有使用成功
linux下直接用
传送文件也可]]></description><category></category></item><item><title><![CDATA[windows安装FFmpeg]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/115111699</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/115111699</guid><author>AI_BN</author><pubDate>Tue, 23 Mar 2021 10:31:24 +0800</pubDate><description><![CDATA[1 安装地址
版本下载
下载最新的版本
2 解压，建立FFmpeg文件夹
将解压的文件拷贝到FFmpeg文件夹中
3 设置路径
将文件夹中bin的路径设置到windows的path路径中。

]]></description><category></category></item><item><title><![CDATA[caffe安装_opencv4.0_aconda cpu python3.8 ubuntu20.04（亲自安装）]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/113715680</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/113715680</guid><author>AI_BN</author><pubDate>Sat, 06 Feb 2021 11:31:53 +0800</pubDate><description><![CDATA[1 安装caffe依赖库
sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
sudo apt-get install --no-install-recommends libboost-all-dev
sudo apt-get install libatlas-base-dev
sudo apt-get install libg]]></description><category></category></item><item><title><![CDATA[linux好用的开发命令]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/112983504</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/112983504</guid><author>AI_BN</author><pubDate>Fri, 22 Jan 2021 14:35:04 +0800</pubDate><description><![CDATA[1 判断两个文件是否一样

md5sum 文件
例如 判断./tmp/test.txt
md5sum /tmp/test.txt会生成文件信息，如果别人使用的是你的文件，就可以判断你们的文件是不是一致了

2导入库
export LD_LIBRARY_PATH=/tmp/app/exec/test/lib/
将库的路径导入进去
3挂载
挂载到其它设备的/home/xxxx路径下
mount -t nfs -o nolock 192.168.17.206:/home/user/dataBase  /home/]]></description><category></category></item><item><title><![CDATA[关于pytorch使用tensorboard以及visdom]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/112837423</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/112837423</guid><author>AI_BN</author><pubDate>Tue, 19 Jan 2021 16:40:29 +0800</pubDate><description><![CDATA[UnicodeDecodeError: ‘utf-8’ codec can’t decode bytes in position 1022-1023:
在python.py开头加上如下：
#!/usr/bin/env Python
coding=utf-8

]]></description><category></category></item><item><title><![CDATA[vs2019的使用]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/111934163</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/111934163</guid><author>AI_BN</author><pubDate>Tue, 29 Dec 2020 17:53:10 +0800</pubDate><description><![CDATA[1根据上图找到工具栏；
2 安装image watch 可以实时观察cv图片

]]></description><category></category></item><item><title><![CDATA[论文解读以及翻译-语义分割]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/111569692</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/111569692</guid><author>AI_BN</author><pubDate>Wed, 23 Dec 2020 00:24:10 +0800</pubDate><description><![CDATA[HIERARCHICAL MULTI-SCALE ATTENTION FOR SEMANTIC SEGMENTATION
层次多尺度注意力机制的语义分割
ABSTRACT
摘要
Multi-scale inference is commonly used to improve the results of semantic segmentation。
多尺度推理是提高语义分割的常见的方式。
Multiple images scales are passed through a network and the]]></description><category></category></item><item><title><![CDATA[himix100 himix200 hisiv300安装]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/111312325</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/111312325</guid><author>AI_BN</author><pubDate>Thu, 17 Dec 2020 11:53:33 +0800</pubDate><description><![CDATA[前提有虚拟机：https://mirrors.tuna.tsinghua.edu.cn/help/ubuntu/
安装himix200
1 下载交叉编译工具：arm-himix200-linux
2 安装交叉环境
tar -zxvf arm-himix200-linux.tgz
cd arm-himix200-linux
sudo ./arm-himix200-linux.install
3 添加环境变量：
export path=/opt/hisi-linux/x86-arm/arm-himix200-l]]></description><category></category></item><item><title><![CDATA[python系列]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/111092620</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/111092620</guid><author>AI_BN</author><pubDate>Sun, 13 Dec 2020 11:33:54 +0800</pubDate><description><![CDATA[1 关于符号的解释
“/”，这是传统的除法，3/2=1.5
“//”，在python中，这个叫“地板除”，3//2=1
“%”，这个是取模操作，也就是区余数，4%2=0，5%2=1

]]></description><category></category></item><item><title><![CDATA[np系列]]></title><link>https://blog.youkuaiyun.com/AI_BN/article/details/111059341</link><guid>https://blog.youkuaiyun.com/AI_BN/article/details/111059341</guid><author>AI_BN</author><pubDate>Sat, 12 Dec 2020 13:26:26 +0800</pubDate><description><![CDATA[1 numpy常用函数
np.arange(n): 元素从0到n-1的ndarray类型
np.ones(shape): 生成全1
np.ones(shape): 生成全1
np.zeros((shape)， ddtype = np.int32) ： 生成int32型的全0
np.full(shape, val): 生成全为val
np.eye(n) : 生成单位矩阵
np.ones_like(a) : 按数组a的形状生成全1的数组
np.zeros_like(a): 同理
np.full_like (a,]]></description><category></category></item></channel></rss>