问题一：libunwind.so.8: undefined reference

//usr/lib/x86_64-linux-gnu/libunwind.so.8: undefined reference to `lzma_index_end@XZ_5.0'

//usr/lib/x86_64-linux-gnu/libunwind.so.8: undefined reference to `lzma_index_size@XZ_5.0'

//usr/lib/x86_64-linux-gnu/libunwind.so.8: undefined reference to `lzma_index_buffer_decode@XZ_5.0'

//usr/lib/x86_64-linux-gnu/libunwind.so.8: undefined reference to `lzma_stream_footer_decode@XZ_5.0'

//usr/lib/x86_64-linux-gnu/libunwind.so.8: undefined reference to `lzma_stream_buffer_decode@XZ_5.0'

//usr/lib/x86_64-linux-gnu/libunwind.so.8: undefined reference to `lzma_index_uncompressed_size@XZ_5.0'

解决方法：

ldconfig -p | grep lzma

liblzma.so.5 (libc6,x86-64) => /lib/x86_64-linux-gnu/liblzma.so.5

liblzma.so.5 (libc6) => /lib/i386-linux-gnu/liblzma.so.5

liblzma.so (libc6,x86-64) => /usr/lib/x86_64-linux-gnu/liblzma.so

export LD_LIBRARY_PATH=/lib/x86_64-linux-gnu/

问题二：

libcudart.so.7.5: cannot open shared object file: No such file or directory

解决方法：

1）管理员直接加入

32-bit：sudo ldconfig /usr/local/cuda/lib

64-bit：sudo ldconfig /usr/local/cuda/lib64

2）服务器加入自己路径下的设置中

首先打开bashrc文件：gedit ~/.bashrc

在文件末尾加入环境变量，如64-bit：export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64"

保存文件退出，在终端保存加载设置：soucre ~/.bashrc

3）临时性加入，下次登录后需要重新加入

export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:/usr/local/cuda/lib64"

问题三：

libhdf5_hl.so.10: cannot open shared object file: No such file or directory

我用anaconda2

解决方法：

export LD_LIBRARY_PATH=/path/to/your/anaconda2/lib:$LD_LIBRARY_PATH

我的是

export LD_LIBRARY_PATH=/home/k/anaconda2/lib:$LD_LIBRARY_PATH

问题四:

fatal error: caffe/proto/caffe.pb.h: No such file or directory #include "caffe/proto/caffe.pb.h"

解决方法:

# In the directory you installed Caffe to

protoc src/caffe/proto/caffe.proto --cpp_out=.

mkdir include/caffe/proto

mv src/caffe/proto/caffe.pb.h include/caffe/proto

问题五:

error while loading shared libraries: libprotoc.so.9: cannot open shared object file: No such file or directory

解决方法:

export LD_LIBRARY_PATH=/usr/local/lib/

问题六:undefined reference to `caffe::db::GetDB(std::string const&)'

collect2: error: ld returned 1 exit status

make: *** [.build_release/tools/extract_features.bin] Error 1

make: *** Waiting for unfinished jobs....

.build_release/tools/compute_image_mean.o: In function `main':

compute_image_mean.cpp:(.text.startup+0x1b0): undefined reference to `caffe::db::GetDB(std::string const&)'

collect2: error: ld returned 1 exit status

make: *** [.build_release/tools/compute_image_mean.bin] Error 1

.build_release/tools/convert_imageset.o: In function `main':

convert_imageset.cpp:(.text.startup+0x625): undefined reference to `caffe::db::GetDB(std::string const&)'

collect2: error: ld returned 1 exit status

make: *** [.build_release/tools/convert_imageset.bin] Error 1

解决方法清除/usr/lib/libcaffe.so或清除/usr/local/lib/libcaffe.so

问题七:

lib/libcaffe.so: undefined reference to `cudnnSetLRNDescriptor'

lib/libcaffe.so: undefined reference to `cudnnGetConvolutionBackwardDataWorkspaceSize'

lib/libcaffe.so: undefined reference to `cudnnSetFilter4dDescriptor'

lib/libcaffe.so: undefined reference to `cudnnGetConvolutionForwardAlgorithm'

lib/libcaffe.so: undefined reference to `cudnnConvolutionBackwardData_v3'

lib/libcaffe.so: undefined reference to `cudnnGetConvolutionBackwardFilterWorkspaceSize'

lib/libcaffe.so: undefined reference to `cudnnCreateConvolutionDescriptor'

lib/libcaffe.so: undefined reference to `cudnnDestroyConvolutionDescriptor'

lib/libcaffe.so: undefined reference to `cudnnSoftmaxBackward'

lib/libcaffe.so: undefined reference to `cudnnConvolutionForward'

lib/libcaffe.so: undefined reference to `cudnnLRNCrossChannelBackward'

lib/libcaffe.so: undefined reference to `curandGenerateUniform'

lib/libcaffe.so: undefined reference to `cudnnPoolingBackward'

lib/libcaffe.so: undefined reference to `cudnnCreateTensorDescriptor'

lib/libcaffe.so: undefined reference to `cublasDdot_v2'

lib/libcaffe.so: undefined reference to `curandGenerateUniformDouble'

lib/libcaffe.so: undefined reference to `cudnnCreatePoolingDescriptor'

lib/libcaffe.so: undefined reference to `cudnnDestroyFilterDescriptor'

lib/libcaffe.so: undefined reference to `cudnnActivationBackward'

lib/libcaffe.so: undefined reference to `cudnnDestroyTensorDescriptor'

lib/libcaffe.so: undefined reference to `cublasDaxpy_v2'

lib/libcaffe.so: undefined reference to `cudnnCreateLRNDescriptor'

lib/libcaffe.so: undefined reference to `curandCreateGenerator'

lib/libcaffe.so: undefined reference to `cudnnGetConvolutionForwardWorkspaceSize'

lib/libcaffe.so: undefined reference to `cublasDscal_v2'

lib/libcaffe.so: undefined reference to `cublasScopy_v2'

lib/libcaffe.so: undefined reference to `cudnnAddTensor'

lib/libcaffe.so: undefined reference to `cudnnDestroyPoolingDescriptor'

lib/libcaffe.so: undefined reference to `cudnnSoftmaxForward'

lib/libcaffe.so: undefined reference to `cudnnActivationForward'

lib/libcaffe.so: undefined reference to `cudnnPoolingForward'

lib/libcaffe.so: undefined reference to `cublasSgemv_v2'

lib/libcaffe.so: undefined reference to `cudnnSetTensor4dDescriptorEx'

lib/libcaffe.so: undefined reference to `curandSetPseudoRandomGeneratorSeed'

lib/libcaffe.so: undefined reference to `cublasSdot_v2'

lib/libcaffe.so: undefined reference to `cudnnSetPooling2dDescriptor'

lib/libcaffe.so: undefined reference to `curandDestroyGenerator'

lib/libcaffe.so: undefined reference to `cudnnConvolutionBackwardBias'

lib/libcaffe.so: undefined reference to `cublasDcopy_v2'

lib/libcaffe.so: undefined reference to `cudnnGetConvolutionBackwardFilterAlgorithm'

lib/libcaffe.so: undefined reference to `cublasDestroy_v2'

lib/libcaffe.so: undefined reference to `curandGenerateNormalDouble'

lib/libcaffe.so: undefined reference to `cublasSgemm_v2'

lib/libcaffe.so: undefined reference to `cudnnCreateFilterDescriptor'

lib/libcaffe.so: undefined reference to `cublasDgemv_v2'

lib/libcaffe.so: undefined reference to `cublasDasum_v2'

lib/libcaffe.so: undefined reference to `cudnnSetStream'

lib/libcaffe.so: undefined reference to `cudnnLRNCrossChannelForward'

lib/libcaffe.so: undefined reference to `cudnnConvolutionBackwardFilter_v3'

lib/libcaffe.so: undefined reference to `curandSetGeneratorOffset'

lib/libcaffe.so: undefined reference to `cublasSaxpy_v2'

lib/libcaffe.so: undefined reference to `cudnnDestroy'

lib/libcaffe.so: undefined reference to `cublasDgemm_v2'

lib/libcaffe.so: undefined reference to `curandGenerate'

lib/libcaffe.so: undefined reference to `cublasSscal_v2'

lib/libcaffe.so: undefined reference to `cudnnSetConvolution2dDescriptor'

lib/libcaffe.so: undefined reference to `cublasSasum_v2'

lib/libcaffe.so: undefined reference to `cublasCreate_v2'

lib/libcaffe.so: undefined reference to `curandGenerateNormal'

lib/libcaffe.so: undefined reference to `cudnnGetConvolutionBackwardDataAlgorithm'

lib/libcaffe.so: undefined reference to `cudnnDivisiveNormalizationBackward'

lib/libcaffe.so: undefined reference to `cudnnDivisiveNormalizationForward'

lib/libcaffe.so: undefined reference to `cudnnCreate'

解决方法:

export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

问题七:

make pycaffe错误:

CXX/LD -o python/caffe/_caffe.so python/caffe/_caffe.cpp

python/caffe/_caffe.cpp:1:52: fatal error: Python.h: No such file or directory

#include // NOLINT(build/include_alpha)

compilation terminated.

make: *** [python/caffe/_caffe.so] Error 1

解决方法：

Makefile.config中anaconda路径不对

问题八:

如果发现编译的时候没有找到cudnn相关的头文件，那么有可能是用了低版本的cudnn，特别是cudnnv4，下载cudnnv5和新版caffe然后把include和lib分别放到cuda响应的include和lib64上然后重新编译。

问题九：

undefined symbol: _ZN5caffe13CuDNNLCNLayerIfED1Ev

找不到libcaffe相关库文件

export LD_LIBRARY_PATH到你编译后使用的libcaffe.so所在位置

wenti10

error while loading shared libraries: libmkl_rt.so: cannot open shared object file: No such file or directory

解决方法：/etc/ld.so.conf.d/ 文件夹下增加了一个文件命名为intel_mkl_settings.conf。然后在里面添加MKL库路径

/opt/intel/mkl/lib/intel64

wenti11、error while loading shared libraries: libcudart.so.5.5: cannot open shared object file: No such file or directory

解决方法：

32-bit: sudo ldconfig /usr/local/cuda/lib

64-bit: sudo ldconfig /usr/local/cuda/lib64

Q8:make: Nothing to be done for `all' 解决方法

1.这句提示是说明你已经编译好了，而且没有对代码进行任何改动。

若想重新编译，可以先删除以前编译产生的目标文件：
make clean
然后再
make

2.出现这种情况解决方法：

a.make clean 清除安装时留下的文件

b.在运行一下ldconfig

c.再make 可运行出结果

Q9:使用caffe的python接口时,很多人都会遇到如下问题:

from ._caffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, \
ImportError: libcaffe.so.1.0.0-rc3: cannot open shared object file: No such file or directory2

这个错误目前非常普遍,在Google上也可以搜索到非常多的讨论,大多数人提出的解决方法是把caffe-master/python和caffe-master/build/lib目录添加到~/.bashrc中,但是问题在于需要修改bashrc,可能会带来很多意想不到的问题.

造成这个问题的主要原因在于使用make进行编译,导致import caffe时找不到静态链接库.

现给出两种比较安全的解决方案:
(1) 在python 代码中加入如下代码:

import os, sys
old_path = os.getcwd()
caffe_root = "caffe-master"
os.chdir(caffe_root + "/python")
import caffe
os.chdir(old_path)

这部分代码就是先切换到caffe的python目录,然后导入caffe库,然后在切换回原来目录.

(2) 其实找不到libcaffe.so.1.0.0-rc3主要原因在于使用了make进行编译,使得caffe的python接口只能在这个目录下面才能找得到这个库.解决办法是使用cmake进行编译:

cd caffe-mastermkdir build
   cd buildcmake ..

这样编译后如果要在python中使用caffe,只需要将caffe的python目录添加到sys.path之中即可:

import oscaffe_root = "caffe-master"
   os.path.append(caffe_root + "/python")

学习：Ubuntu14.04编译caffe问题记录相关推荐

android swstnw cn,Android应用开发之ubuntu14.04编译ijkplayer备忘
本文将带你了解Android应用开发之ubuntu14.04编译ijkplayer备忘,希望本文对大家学Android有所帮助 1. i|VIhzkKCf#3UUnstall gi[e-RkYJ4 ...
基于Ubuntu14.04搭建caffe训练环境
各种教学视频或文档资料 +QQ:635992897 一.基础环境配置 0.apt-get update 1.sshd安装 sudo apt-get install openssh-server ...
Ubuntu14.04 + Matlab2014a + caffe + cuda + cudnn环境搭建
为什么80%的码农都做不了架构师?>>> 安装Caffe环境前,请查好各个软件版本的相互匹配资料,确定好所有软件版本后,再开始安装. 一.安装gcc4.6 Ubuntu14.0 ...
ubuntu14.04+gtx1080+caffe+cuda8.0+matlab2015b以及遇到的问题
一开始是只菜鸟,linux是啥都闹不清,因为实验室要用caffe研究深度学习,然后自己折腾这玩意不知道折腾了有多久,重装系统不下20次!!!照着人家的教程来,总也遇见人家遇不见的问题,好多时候解决不了 ...
搭建Ubuntu14.04编译环境并下载源代码（一）
要想在Ubuntu下编译Android源代码,首先要搭建好开发环境參考链接(官方): 环境:http://source.android.com/source/initializing.html 下载 ...
ubuntu14.04上Trac配置记录
系统环境:ubuntu14.04 (并假设Apache2服务可以正常运行) 1. 安装软件: sudo aptitude install trac python-mysqldb 2. 创建数据库 Tr ...
ubuntu14.04配置caffe
ubuntu 14.04 64bit 安装请自行Google安装,并修改源. cuda 7.5 安装 cuda7.5下载,选择ubuntu 14.04的deb包下载安装一些可能的依赖 sudo a ...
Ubuntu 16.04安装Caffe的记录及FCN官方代码的配置
相关内容搜集自官方文档与网络,既无创新性,也不求甚解,我也不了解Caffe,仅仅搭上之后做个记录,方便以后重装安装依赖项 sudo apt-get install libprotobuf-dev l ...
虚拟机ubuntu14.04编译MPI版本NAMD
最近,尝试用ubuntu编译NAMD,由于以前对NAMD完全未接触过,也不懂该专业的相关知识,因此作为一个外行人尝试编译NAMD,这里记录一下粗略过程: 1.首先,需要搭建MPI并行环境:http:/ ...
ubuntu14.04编译安装strongswan
原文链接:https://blog.csdn.net/wwwangdeqqq/article/details/52388625 因为需要在strongswan基础上做些二次开发的东西,需要将自己修改后 ...

学习：Ubuntu14.04编译caffe问题记录

1）管理员直接加入

2）服务器加入自己路径下的设置中

3）临时性加入，下次登录后需要重新加入

学习：Ubuntu14.04编译caffe问题记录相关推荐

最新文章

热门文章