mxnet环境搭建随记
安装mxnet还是遇到不少麻烦事,现在简单记一下,挖个坑,后续详细补充,打算写一下我的mxnet探索之旅。
更新: 具体安装mxnet,cuda,opencv过程已经补坑——点击进入
不知道为什么,在cuda7.5+ubuntu14.04amd64 + opencv2.4.13时,会显卡驱动有问题,无法显示桌面.但是可以正常使用。后续我改为16.04+cuda8.0rc+opencv2.4.13,解决了桌面显示问题且正常使用,但是在想把opencv2.4.13换为opencv3.10时,编译opencv3.10时会报错,关于cudalegacy模块,如下;
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘NppiGraphcutState’ has not been declared
typedef NppStatus (*init_func_t)(NppiSize oSize, NppiGraphcutState** ppStat
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘NppiGraphcutState’ does not name a type
operator NppiGraphcutState*()
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘NppiGraphcutState’ does not name a type
NppiGraphcutState* pState;
^
In file included from /home/ecarx/opencv-3.1./build/modules/cudalegacy/precomp.hpp:::
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp: In constructor ‘{anonymous}::NppiGraphcutStateHandler::NppiGraphcutStateHandler(NppiSize, Npp8u*, {anonymous}::init_func_t)’:
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘pState’ was not declared in this scope
nppSafeCall( func(sznpp, &pState, pDeviceMem) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp: In destructor ‘{anonymous}::NppiGraphcutStateHandler::~NppiGraphcutStateHandler()’:
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘pState’ was not declared in this scope
nppSafeCall( nppiGraphcutFree(pState) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcutFree’ was not declared in this scope
nppSafeCall( nppiGraphcutFree(pState) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp: In function ‘void cv::cuda::graphcut(cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::Stream&)’:
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcutGetSize’ was not declared in this scope
nppSafeCall( nppiGraphcutGetSize(sznpp, &bufsz) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcutInitAlloc’ was not declared in this scope
NppiGraphcutStateHandler state(sznpp, buf.ptr<Npp8u>(), nppiGraphcutInitAll
^
In file included from /home/ecarx/opencv-3.1./build/modules/cudalegacy/precomp.hpp:::
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcut_32s8u’ was not declared in this scope
nsp.step), sznpp, labels.ptr<Npp8u>(), static_cast<int>(labels.step), state) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcut_32f8u’ was not declared in this scope
nsp.step), sznpp, labels.ptr<Npp8u>(), static_cast<int>(labels.step), state) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp: In function ‘void cv::cuda::graphcut(cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::GpuMat&, cv::cuda::Stream&)’:
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcut8GetSize’ was not declared in this scope
nppSafeCall( nppiGraphcut8GetSize(sznpp, &bufsz) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcut8InitAlloc’ was not declared in this scope
NppiGraphcutStateHandler state(sznpp, buf.ptr<Npp8u>(), nppiGraphcut8InitAl
^
In file included from /home/ecarx/opencv-3.1./build/modules/cudalegacy/precomp.hpp:::
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcut8_32s8u’ was not declared in this scope
nsp.step), sznpp, labels.ptr<Npp8u>(), static_cast<int>(labels.step), state) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
/home/ecarx/opencv-3.1./modules/cudalegacy/src/graphcuts.cpp::: error: ‘nppiGraphcut8_32f8u’ was not declared in this scope
nsp.step), sznpp, labels.ptr<Npp8u>(), static_cast<int>(labels.step), state) );
^
/home/ecarx/opencv-3.1./modules/core/include/opencv2/core/private.cuda.hpp::: note: in definition of macro ‘nppSafeCall’
#define nppSafeCall(expr) cv::cuda::checkNppError(expr, __FILE__, __LINE__, CV
^
modules/cudalegacy/CMakeFiles/opencv_cudalegacy.dir/build.make:: recipe for target 'modules/cudalegacy/CMakeFiles/opencv_cudalegacy.dir/src/graphcuts.cpp.o' failed
make[]: *** [modules/cudalegacy/CMakeFiles/opencv_cudalegacy.dir/src/graphcuts.cpp.o] Error
make[]: *** 正在等待未完成的任务....
[ %] Linking CXX executable ../../bin/opencv_perf_photo
CMakeFiles/Makefile2:: recipe for target 'modules/cudalegacy/CMakeFiles/opencv_cudalegacy.dir/all' failed
make[]: *** [modules/cudalegacy/CMakeFiles/opencv_cudalegacy.dir/all] Error
make[]: *** 正在等待未完成的任务....
[ %] Built target opencv_perf_photo
Makefile:: recipe for target 'all' failed
make: *** [all] Error
编译opencv3.10报错
经多次尝试,都不成功。暂时放弃使用,初步猜测是因为,opencv3.10是2015年年末发行的,cuda8.0rc是2016.05发行的,所以本身opencv3.10不支持cuda8.0rc,但是我还没有去github问(opencv2.4.13是2016.05发行,使用正常),后续还会再尝试,到时再更新。
我个人是因为硬件问题,所以环境选ubuntu16.04+cuda8.0rc。如果没这个限制最好选择比最新版本低一个版本,以求稳定。
不得不说,搭建环境很耗时间,环境稳定后,不要轻易改变。
mxnet环境搭建随记的更多相关文章
- HBase环境搭建随记
====软件版本==== jdk:jdk-8u77-linux-x64.tar.gz zookeeper:zookeeper-3.4.6.tar.gz hadoop:hadoop-2.7.4.tar. ...
- ubantu16.04+mxnet +opencv+cuda8.0 环境搭建
ubantu16.04+mxnet +opencv+cuda8.0 环境搭建 建议:环境搭建完成之后,不要更新系统(内核) 转载请注明出处: 微微苏荷 一 我的安装环境 系统:ubuntu16.04 ...
- 痞子衡嵌入式:语音处理工具Jays-PySPEECH诞生记(1)- 环境搭建(Python2.7.14 + PyAudio0.2.11 + Matplotlib2.2.3 + SpeechRecognition3.8.1 + pyttsx3 2.7)
大家好,我是痞子衡,是正经搞技术的痞子.今天痞子衡给大家介绍的是语音处理工具Jays-PySPEECH诞生之环境搭建. 在写Jays-PySPEECH时需要先搭好开发环境,下表列出了开发过程中会用到的 ...
- 痞子衡嵌入式:串口调试工具Jays-PyCOM诞生记(1)- 环境搭建(Python2.7.14 + pySerial3.4 + wxPython4.0.3)
大家好,我是痞子衡,是正经搞技术的痞子.今天痞子衡给大家介绍的是串口调试工具Jays-PyCOM诞生之环境搭建. 在写Jays-PyCOM时需要先搭好开发和调试环境,下表列出了开发过程中会用到的所有软 ...
- windows kafka 环境搭建踩坑记
版本介绍(64位): Windows 10 JDK1.8.0_171 zookeeper-3.4.8/ kafka_2.11-0.10.0.1.tgz 点击链接进行下载 1. JDK安装和环境搭建 自 ...
- 踩坑记:Tensorflow环境搭建
自从上一篇论文投出去,之后就各种事就来了……处理那些乱七八糟的事就是让人心累,在加上师哥们毕业,能帮我的人越来越少了,而要指望你的人呢,越来越多.一想到那些用搜索引擎都搜不到资料的人,蓦地想起邓爷爷说 ...
- mac开发环境爬坑记(搭建php+nginx+mysql+redis+laravel+git+phpstorm)
题外话:前几天,终于以原价一半的价格,将我那台15版mbp在bbs上卖了出去.之所以用了“终于”这个词儿,是我一直迟迟没有下定决心卖掉它,可眼瞅着再不卖掉,又要掉价,况且我的新电脑,也终于下来了. 话 ...
- .NET Core系列 : 1、.NET Core 环境搭建和命令行CLI入门
2016年6月27日.NET Core & ASP.NET Core 1.0在Redhat峰会上正式发布,社区里涌现了很多文章,我也计划写个系列文章,原因是.NET Core的入门门槛相当高, ...
- JavaScript学习(一) —— 环境搭建与JavaScript初探
1.开发环境搭建 本系列教程的开发工具,我们采用HBuilder. 可以去网上下载最新的版本,然后解压一下就能直接用了.学习JavaScript,环境搭建是非常简单的,或者说,只要你有一个浏览器,一个 ...
随机推荐
- PHP中面相对象对象的知识点整理
面向对象 万物皆对象,将构成问题的事务分解到各个对象上,建立对象的目的不是为了完成一个工作,而是为了描述某个事务在解决问题中的行为,更符合人的思维习惯,代码重用性高,可扩展性. ___________ ...
- HDU 1423 最长公共字串+上升子序列
http://acm.hdu.edu.cn/showproblem.php?pid=1423 在前一道题的基础上多了一次筛选 要选出一个最长的递增数列 lower_bound()函数很好用,二分搜索找 ...
- xdebug调试php程序
相关设置 xdebug.default_enable=1 默认是1,当错误出现时,堆栈跟踪会激活.可以在代码中通过xdebug_disable()来关闭它. xdebug.force_display_ ...
- zabbix之Nginx安装
转载自 http://www.ttlsa.com/nginx/nginx-install-on-linux/ Nginx下载 https://pan.baidu.com/s/1qXT54sO
- gjd
#include <cstdio> #include <cstring> #include <malloc.h> #define radix (1u<< ...
- 使用Django——安装
1. 安装 a) 安装python 在http://www.python.org/上下载python 2.7,然后安装,接着将python的安装路径(一般是C:\python27)添加到windows ...
- 给一个正在运行的Docker容器动态添加Volume
给一个正在运行的Docker容器动态添加Volume本文转自:http://dockone.io/article/149 [编者的话]之前有人问我Docker容器启动之后还能否再挂载卷,考虑到mnt命 ...
- iOS GCD 必读推荐,有关于单例使用问题
链接如下:http://www.cocoachina.com/swift/20150129/11057.html 以前只注意使用dispatch_once达到创建单例对象时的线程安全,读了下边这篇文章 ...
- java web 学习 --第一天(Java三级考试)
1.Servlet servlet是运行在web server或 application server端的Java程序,主要用于在服务器端产生动态内容. servlet 在服务器端主要有以下作用 读取 ...
- Divide and conquer:Subset(POJ 3977)
子序列 题目大意:给定一串数字序列,要你从中挑一定个数的数字使这些数字和绝对值最小,求出最小组合数 题目的数字最多35个,一看就是要数字枚举了,但是如果直接枚举,复杂度就是O(2^35)了,显然行不通 ...