Nvidia driver-410
系统自带440驱动要装cuda10,先将驱动全部卸载,再安装410
卸载
sudo apt-get --purge remove nvidia*
sudo apt autoremove
自带驱动屏蔽
在终端输入
lsmod | grep nouveau
如果有输出需要禁用系统自带的 nouveau 驱动↓,没有就跳过
创建一个配置文件
sudo vim /etc/modprobe.d/blacklist-nouveau.conf
在该配置文件中添加如下内容
blacklist nouveau
options nouveau modeset=0
进行更新
sudo update-initramfs -u
然后重启,在终端输入
lsmod | grep nouveau
无输出则成功
安装
sudo apt-get install nvidia-driver-410
检查
lsmod | grep nvidia
输出类似↓就安装成功了
nvidia_uvm 790528 0
nvidia_drm 40960 0
nvidia_modeset 1040384 1 nvidia_drm
nvidia 16633856 2 nvidia_uvm,nvidia_modeset
drm_kms_helper 172032 2 mgag200,nvidia_drm
drm 458752 6 drm_kms_helper,mgag200,nvidia_drm,ttm
ipmi_msghandler 102400 4 ipmi_devintf,ipmi_si,nvidia,ipmi_ssif
也可以nvidia-smi查看驱动版本
cuda9.0
GCC & G++
Ubuntu18.04预装GCC7.3,而CUDA9.0支持GCC6.0以下版本。gcc和g++从自带的7降级到6
查看版本
Gcc --version
G++ --version
降级
sudo apt install gcc-6 g++-6
sudo ln -s /usr/bin/gcc-6 /usr/local/bin/gcc
sudo ln -s /usr/bin/g++-6 /usr/local/bin/g++
cuda9.0
安装
官网下载cuda_9.0.176_384.81_linux-run和cuda_9.0.176.1_linux.run文件,cd到目录下
Chmod +x cuda_9.0.176_384.81_linux-run
./ cuda_9.0.176_384.81_linux-run
除了!!!跳过安装驱动!!!,全部选默认,安装完成
#安装补丁
Chmod +x cuda_9.0.176.1_linux-run
./ cuda_9.0.176.1_linux-run
添加环境变量
sudo vim ~/.bashrc
在结尾添加
#cuda9.0
export PATH=$PATH:/usr/local/cuda/bin/
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64/
重启
检查是否安装成功
1.查看版本
Nvcc -V
正常结果↓
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Sep__1_21:08:03_CDT_2017
Cuda compilation tools, release 9.0, V9.0.176
2.运行example
cd /usr/local/cuda-9.0/samples/1_Utilities/deviceQuery
sudo make
sudo ./deviceQuery
输出GPU相关信息就对啦,一定要sudo
Cudnn7
官网下载cudnn-9.0-linux-x64-v7.6.4.38.tgz,cd到目录里
tar -zxvf cudnn-9.0-linux-x64-v7.6.4.38.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
sudo chmod a+r /usr/local/cuda/lib64/libcudnn*
miniconda3
官网下载Miniconda3-latest-Linux-x86_64.sh,cd到目录里
chmod +x Miniconda3-latest-Linux-x86_64.sh
./ Miniconda3-latest-Linux-x86_64.sh
RefineDet
Base安装(成功)
编译依赖项
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 libopenblas-dev liblapack-dev libatlas-base-dev
sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
装python2的依赖项:
sudo apt-get install python-pip
sudo apt-get install python-scipy
sudo apt-get install python-matplotlib
sudo apt-get install python-skimage
sudo apt-get install python-dev
sudo apt-get install python-numpy
sudo apt-get install opencv-python
安装RefineDet
Cd到 $RefineDet_ROOT.
git clone https://github.com/sfzhang15/RefineDet.git
cp Makefile.config.example Makefile.config
#修改Makefile.config (opencv3那行取消注释)
make all -j
make py
要用python2运行
cd ****/refinedet/python
python2
>>import caffe
不报错就是成功了,如果报cublas的错↓是因为没安装cuda补丁(见安装cuda部分↑)
F1121 15:56:26.234781 27262 math_functions.cu:26] Check failed: status == CUBLAS_STATUS_SUCCESS (13 vs. 0) CUBLAS_STATUS_EXECUTION_FAILED
*** Check failure stack trace: ***)
虚拟环境(失败待解决)
创建2.7环境
安装RefineDet后import报错
>>> import caffe
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "caffe/__init__.py", line 1, in <module>
from .pycaffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, RMSPropSolver, AdaDeltaSolver, AdamSolver
File "caffe/pycaffe.py", line 13, in <module>
from ._caffe import Net, SGDSolver, NesterovSolver, AdaGradSolver, \
ImportError: /usr/lib/libgdal.so.20: undefined symbol: sqlite3_column_table_name
Update sqlite无效,未解决。