반응형

Follow this guide from Nvidia

docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html

1. Verify having CUDA-capable machine

$ lspci | grep -i nvidia

2. Check Linux version

uname -m && cat /etc/*release

3. Check gcc compiler

gcc --version

 

Install Cuda 10.1

developer.nvidia.com/cuda-10.1-download-archive-base?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1804&target_type=deblocal

 

 

Broken Package

askubuntu.com/questions/1026739/dpkg-error-processing-archive-var-cache-apt-archives-cuda-cublas-9-1-9-1-85-3

 

cuDNN

Download files following :

developer.nvidia.com/rdp/cudnn-download

and follow this process

docs.nvidia.com/deeplearning/cudnn/install-guide/index.html#installcuda

Download Library for cuDNN

developer.nvidia.com/compute/machine-learning/cudnn/secure/8.0.2.39/10.1_20200724/cudnn-10.1-linux-x64-v8.0.2.39.tgz

Extract above file and copy to a kind of folder [1]

And if you just run deep-learning model, download runtime library. but, want to check proper operation, download developer library

developer.nvidia.com/compute/machine-learning/cudnn/secure/8.0.2.39/10.1_20200724/Ubuntu18_04-x64/libcudnn8-dev_8.0.2.39-1%2Bcuda10.1_amd64.deb

if you contact 'libcudnn.so.8 is not symbolic link', following below

stackoverrun.com/ko/q/9962800

 

 

Ref

[1] idchowto.com/?p=52738

leesh0523.tistory.com/entry/NDIVIA-Driver-CUDA-cuDNN-%EC%84%A4%EC%B9%98%ED%95%98%EA%B8%B0

www.pyimagesearch.com/2019/01/30/ubuntu-18-04-install-tensorflow-and-keras-for-deep-learning/

728x90

+ Recent posts