Ubuntu16.04
下载
CUDA : CUDA Toolkit 8.0 GA2 (Feb 2017), Online Documentation
cuDNN：cuDNN v7.1.3 Library for Linux
一、安装Nvidia驱动
笔记本 GTX 1050ti
Nvidia 官网
因为我是在笔记上配的环境所以产品系列选的是(Notebooks)
屏蔽开源驱动nouveau
sudo vim /etc/modprobe.d/blacklist-nouveau.conf
内容为
blacklist nouveau
options nouveau modeset=0
再更新
sudo update-initramfs -u
安装驱动
sudo sh ./NVIDIA-Linux-x86_64-450.57.run -no-x-check -no-nouveau-check -no-opengl-files
检查是否安装成功
nvidia-smi
在~/.bashrc添加环境变量
export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH
export LD_LIBRARY_PATH=/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH
二、安装CUDA8.0
https://developer.nvidia.com/cuda-toolkit-archive
sudo dpkg -i cuda-repo-ubuntu1604-8-0-local-ga2_8.0.61-1_amd64.deb
sudo apt-get update
sudo apt-get install cuda -y
检查是否安装成功
安装cuda sample
cd /usr/local/cuda-11.0/samples
sudo make -j32
编译完成之后
cd bin/x86_64/linux/release/
sudo ./deviceQuery
在~/.bashrc中添加环境变量
# CUDA
export PATH=/usr/local/cuda-8.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
三、安装cuDNN
https://developer.nvidia.com/rdp/cudnn-archive
tar xzvf cudnn-8.0-linux-x64-v7.1.tgz
cd cuda/
sudo cp lib64/lib* /usr/local/cuda/lib64/
sudo cp include/cudnn.h /usr/local/cuda/include/
sudo chmod +r libcudnn.so.7.1.3
sudo ln -sf libcudnn.so.7.1.3 libcudnn.so.7
sudo ln -sf libcudnn.so.7 libcudnn.so
sudo ldconfig
四、安装BLAS
sudo apt-get install libopenblas-dev
最后需要修改Caffe 的Makefile.config
BLAS:=open
五、C++ 依赖项
sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libboost-all-dev libhdf5-serial-dev libgflags-dev libgoogle-glog-dev liblmdb-dev protobuf-compiler protobuf-c-compiler protobuf-compiler -y
六、python 依赖项
sudo apt-get install python-numpy python-scipy python-matplotlib python-sklearn python-skimage python-h5py python-protobuf python-leveldb python-networkx python-nose python-pandas python-gflags cython ipython
七、Makefile.config
cd caffe-master
cp Makefile.config.example Makefile.config
启用cuDNN，去掉#
USE_CUDNN := 1
设置BLAS
BLAS := open
配置引用文件
INCLUDE_DIRS := $(PYTHON_INCLUDE) /usr/local/include /usr/lib/x86_64-linux-gnu/hdf5/serial/include
LIBRARY_DIRS := $(PYTHON_LIB) /usr/local/lib /usr/lib /usr/lib/x86_64-linux-gnu/hdf5/serial
配置python路径
PYTHON_LIB := /usr/local/lib
八、编译Caffe
make all -j 32
make test -j 32
make runtest -j 32
安装成功