opencv安装
查看opencv版本:
pkg-config opencv --modversion
opencv2.4.9:
opencv3.4.5
安装:ubuntu安装opencv的正确方法_陶陶name-CSDN博客_ubuntu安装opencv
nvida安装
首先我们需要添加源,
sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt update
然后检查可以安装的驱动版本,
ubuntu-drivers devices
选择最合适的版本安装即可,或者自动安装
sudo ubuntu-drivers autoinstall/sudo apt install nvidia-driver-XXX
最后,重启Ubuntu系统,输入nvidia-smi查看驱动信息。
如何变更呢?(可能由于版本问题需要变更)
点击ubuntu中software&&updates
在additional drivers中选择对应的版本进行替换
可以利用sudo apt install nvidia-driver-470来测试,如果没有新安装的,应该就好了,然后重启电脑。
sudo reboot
nvidia-smi(出现表格,有表格再做后续的)
cuda安装(最后是cuda9.2.148,cudnn7.6.2版本)
Ubuntu18.04+显卡驱动安装+CUDA9.1安装+cdDNN7.0.5安装+Anaconda3安装+tensorflow(GPU)安装_鬼刀-CSDN博客
进入nvidia开发者网站的CUDA下载页面 https://developer.nvidia.com/cuda-toolkit-archive
如果已经安装了,如何卸载cuda?
cuda10.0及以下的卸载:
1. cd /usr/local/cuda-xx.x/bin/
2. sudo ./uninstall_cuda_xx.x.pl
3. sudo rm -rf /usr/local/cuda-xx.x
cuda10.1及以上的卸载:
1. cd /usr/local/cuda-xx.x/bin/
2. sudo ./cuda-uninstaller
3. sudo rm -rf /usr/local/cuda-xx.x
ubuntu18.4下cuda卸载 - CVerTian - 博客园
ubuntu16.04/18.04安装卸载cuda10.0/10.1和cudnn图文说明_mbytes的博客-CSDN博客
CUDA安装,同时安装其补丁(3个)
在cuda下载目录下运行终端输入命令安装Base Installer
sudo sh cuda_9.1.85_387.26_linux.run+补丁
10.1
之前已经安装了Nvidia的显卡驱动,这里不安装driver
安装完成
9.2
sudo sh cuda_9.2.148_396.37_linux.run cuda_9.2.148.1_linux.run
首先是说明,按回车知道100%,然后accept
安装成功
安装cudnn
下载链接: NVIDIA cuDNN | NVIDIA Developer
下载好后直接命令行解压,然后复制 lib64 和 include 文件夹到 /usr/local/cuda-10.0,命令如下:
tar -zxvf cudnn-10.0-linux-x64-v7.5.0.56.tgz
sudo find / -name cudnn.h(找文件所在路径)
#10.0
sudo cp /home/heixiazi/add/cuda/include/cudnn.h /usr/local/cuda-10.0/include
sudo cp /home/heixiazi/add/cuda/lib64/libcudnn* /usr/local/cuda-10.0/lib64
sudo chmod a+r /usr/local/cuda-10.0/include/cudnn.h /usr/local/cuda-10.0/lib64/libcudnn*
#10.1
sudo cp /home/heixiazi/add/cuda/include/cudnn.h /usr/local/cuda-10.1/include
sudo cp /home/heixiazi/add/cuda/lib64/libcudnn* /usr/local/cuda-10.1/lib64
sudo chmod a+r /usr/local/cuda-10.1/include/cudnn.h /usr/local/cuda-10.1/lib64/libcudnn*
#9.2
sudo cp /home/heixiazi/add/cuda/include/cudnn.h /usr/local/cuda-9.2/include
sudo cp /home/heixiazi/add/cuda/lib64/libcudnn* /usr/local/cuda-9.2/lib64
sudo chmod a+r /usr/local/cuda-9.2/include/cudnn.h /usr/local/cuda-9.2/lib64/libcudnn*
添加环境变量
sudo gedit ~/.bashrc
在文件末尾添加
#10.1
export PATH="/usr/local/cuda-10.1/bin:$PATH"
export LD_LIBRARY_PATH="/usr/lcoal/cuda-10.1/lib64:$LD_LIBRARY_PATH"
#9.2
export PATH="/usr/local/cuda-9.2/bin:$PATH"
export LD_LIBRARY_PATH="/usr/lcoal/cuda-9.2/lib64:$LD_LIBRARY_PATH"
最后使其生效
source ~/.bashrc
终端输入
#10.1
cd /usr/local/cuda-10.1/samples/1_Utilities/deviceQuery
sudo make
./deviceQuery
#9.2
cd /usr/local/cuda-9.2/samples/1_Utilities/deviceQuery
sudo make
./deviceQuery
查看cuda版本一:
nvcc --version
Command 'nvcc' not found, but can be installed with:
sudo apt install nvidia-cuda-toolkit
han@han-System-Product-Name:~$ sudo apt install nvidia-cuda-toolkit(安装好后就可以查了)
nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Tue_Jun_12_23:07:04_CDT_2018
Cuda compilation tools, release 9.2, V9.2.148
查看cuda的版本二:
cat /usr/local/cuda/version.txt
查看cudnn 版本:
cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2
anaconda
安装:https://blog.csdn.net/gaoyu1253401563/article/details/82808269
卸载 安装在install_software/anaconda3,所以就是
sudo rm -rf install_software/anaconda3
环境配置:
# 将anaconda的bin目录加入PATH,根据版本不同,也可能是~/anaconda3/bin
echo 'export PATH=/home/heixiazi/anaconda3/bin:$PATH' >> ~/.bashrc
# 更新bashrc以立即生效
source ~/.bashrc
tendorflow环境:安装
一个tensorflow环境:python-3.6.13
source activate tensorflow
source deactivate
如何看cuda与pytorch有没有装好
(https://blog.csdn.net/nyist_yangguang/article/details/115011602)
import torch
# 我在python的命令环境测试该命令 返回是True的但是在run torch程序的时候报错
torch.cuda.is_available()
#会显示一句话
import torch
torch.zeros(1).cuda()