【Ubuntu】Ubuntu16.04安装cuda9.0+caffe

首先更新下系统:

sudo apt-get update
sudo apt-get upgrade -y
sudo apt-get dist-upgrade -y

第一步我们要安装和cuda版本匹配的Nvidia驱动,加假如电脑安装了旧版本的驱动,使用如下命令卸载它:

sudo apt-get remove nvidia*
sudo apt-get autoremove

安装依赖项:

sudo apt-get install dkms build-essential linux-headers-generic apt-show-versions

接着要将系统默认的'nouveau'驱动加入黑名单:

sudo gedit /etc/modprobe.d/blacklist-nouveau.conf

将以下内容添加到该文件中:

blacklist nouveau
blacklist lbm-nouveau
options nouveau modeset=0
alias nouveau off
alias lbm-nouveau off

输入以下命令以关闭nouveau内核:

echo options nouveau modeset=0 | sudo tee -a /etc/modprobe.d/nouveau-kms.conf
sudo update-initramfs -u

下载和cuda匹配的驱动,并赋予其执行权限:

chmod +x NVIDIA-Linux******************.run

继续安装依赖项:

sudo apt-add-repository ppa:ubuntu-x-swat/x-updates #if it fails, you may need to repeat the operation.
sudo apt-get update
sudo apt-get install nvidia-current -y
sudo apt-get upgrade -y

注销退出图形界面,Ctrl+Alt+F1进入命令行登录,输入service lightdm stop关闭图形管理器,然后安装nvidia驱动:

sudo ./NVIDIA-Linux******************.run
  • You may have a warning "The distribution-provided pre-install script failed! Are you sure you want to continue ?" => You can continue.
  • Register the kernel module with DKMS => YES !
  • If you don't have 32bits compatibility you will get a warning => OK !
  • Run NVIDIA-Xconfig => YES !

重启系统:

sudo reboot

使用nvidia-smi命令检查安装是否成功:

root@machine: nvidia-smi

+------------------------------------------------------+
| NVIDIA-SMI ***.**    Driver Version: ***.**          |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce GTX TIT...  Off  | 0000:05:00.0     Off |                  N/A |
| 22%   59C    P0    74W / 250W |     23MiB / 12284MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   1  GeForce GTX TIT...  Off  | 0000:06:00.0     Off |                  N/A |
| 22%   60C    P0    72W / 250W |     23MiB / 12284MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+
|   2  GeForce GTX TIT...  Off  | 0000:09:00.0     Off |                  N/A |
| 22%   54C    P0    53W / 250W |     23MiB / 12284MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID  Type  Process name                               Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

第二步,安装cuda,在英伟达官网下载Linux-x86_64-Ubuntu-16.04-deb(local),使用以下命令进行安装:

sudo dpkg -i cuda-repo-ubuntu***************.deb

sudo apt-get update
sudo apt-get install cuda

# setting the environment variables so CUDA will be found
echo -e "\nexport PATH=/usr/local/cuda-9.0/bin${PATH:+:${PATH}} " >> .bashrc
echo -e "\nexport LD_LIBRARY_PATH=/usr/local/cuda-9.0/lib64\${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}" >> .bashrc

检查是否安装成功:

root@machine: apt-show-versions cuda
**cuda:amd64/unknown 9.0-** uptodate**

root@machine: cat /proc/driver/nvidia/version
NVRM version: NVIDIA UNIX x86_64 Kernel Module  ***.**  ************************
GCC version:  gcc version *.*.* (Ubuntu ****************)

运行cuda_samples,如果运行成功就大功告成:

cd /usr/local/cuda-9.0/samples/5_Simulations/nbody
sudo make
./nbody

第三步下载cuDNN v7 Library for Linux安装cuDNN v7加速库:

tar -xzvf cudnn-9.0-linux-x64-v7.tgz

sudo cp cuda/include/cudnn.h /usr/local/cuda/include
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*

使之生效

sudo ldconfig –v

建立库链接:

cd /usr/local/cuda/lib64

sudo ln -s libcudnn.so.7.0.xx  libcudnn.so.7

sudo ln -s libcudnn .so.7  libcudnn.so

最后安装配置caffe框架:

下载caffe源码到本地

安装依赖项:

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 -y libgflags-dev libgoogle-glog-dev liblmdb-dev
sudo apt-get install -y python-numpy python-scipy python-matplotlib python-sklearn python-skimage python-h5py python-protobuf python-leveldb python-networkx python-nose python-pandas python-gflagssudo apt-get install libopenblas-dev libatlas-base-devsudo apt-get install the python-dev

编译前需根据自己需要修改配置文件Makefile.config,进入caffe根目录,首先将示例文件复制一份:

cp Makefile.config.example Makefile.config
使用gedit打开进行修改

编译时可能会遇到的错误:

1..src/caffe/layers/hdf5_data_layer.cpp:13:18: fatal error: hdf5.h: 没有那个文件或目录
compilation terminated.
Makefile:581: recipe for target '.build_release/src/caffe/layers/hdf5_data_layer.o' failed
make: *** [.build_release/src/caffe/layers/hdf5_data_layer.o] Error 1

使用whereis hdf5定位以下hdf5的文件夹,将其路径添加至Makefile.config中,将:

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include
修改为:

INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/include/hdf5/serial/

另外在Makefile中将:

LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_hl hdf5

修改为:

LIBRARIES += glog gflags protobuf boost_system boost_filesystem m hdf5_serial_hl hdf5_serial

修改完毕后开始编译:

sudo make clean
make all
make test
make runtest

运行MNIST例子:

进入caffe/data/mnist执行./get_mnist.sh下载MNIST数据集

运行在caffe根目录运行./examples/mnist/create_mnist.sh将数据集转换为caffe能处理的格式

$CAFFE_ROOT/examples/mnist/lenet_solver.prototxt文件中修改训练使用CPU还是GPU

在caffe根目录下运行./examples/mnist/train_lenet.sh来运行MNIST示例
阅读更多
想对作者说点什么?

博主推荐

换一批

没有更多推荐了,返回首页