How to install OpenCV Nvidia CUDA driver on Ubuntu 18.04 for Tensorflow.

Purpose: Easily setting up OpenCV CUDA ready environment for Deep Neural Network accelerator

This demonstration has been tested on Linux Kernel Ubuntu 18.04 on Windows 10 x64 and pure Ubuntu 18.04

OpenCV Nvidia CUDA GPU driver installation

Ubuntu 18.04 on Windows 10 64 bit
https://discuss.openai.com/t/installing-openai-gym-universe-on-windows/2092

Or

Clean installation of Ubuntu 18.04

Requirement

Python 3.6
OpenCV 3.4.0
Tensorflow 1.9

Instalation

sudo su -
apt-get update
apt-get install -y python3-pip
pip3 install --upgrade pip
sudo apt-get install python-numpy
sudo apt-get install build-essential cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev
sudo apt-get install python3.6-dev
python3.6-config --includes
#The output would be
-I/usr/include/python3.6m -I/usr/include/python3.6m
cd /usr/include/python3.6m
mv pyconfig.h pyconfig.h.bak
cp /usr/include/x86_64-linux-gnu/python3.5m/pyconfig.h /usr/include/python3.5m/
cd /root
mkdir OpenCV-tmp
cd OpenCV-tmp
git clone https://github.com/Itseez/opencv.git
cd opencv
git checkout 3.4.0
cd ../OpenCV-tmp
git clone https://github.com/Itseez/opencv_contrib.git
cd opencv_contrib
git checkout 3.4.0

cd ../opencv
$ mkdir build
$ cd build
$ cmake -D CMAKE_BUILD_TYPE=RELEASE \
    -D CMAKE_INSTALL_PREFIX=/usr/local \
    -D INSTALL_C_EXAMPLES=ON \
    -D INSTALL_PYTHON_EXAMPLES=ON \
    -D OPENCV_EXTRA_MODULES_PATH=/root/OpenCV-tmp/opencv_contrib/modules \
    -D BUILD_EXAMPLES=ON ..

make -j $(nproc --all)
sudo make install
sudo ldconfig

Checking

python3
import cv2
cv2.__version__

(Optional)Install tqdm for model evaluation

pip3 install tqdm

install numpy matplotlib pandas

pip3 install numpy
pip3 install matplotlib
pip3 install pandas

Install tensorflow (CPU version)

pip3 install tensorflow

Install tensorflow-gpu (GPU version) on Ubuntu 18.04

sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt update
sudo apt install nvidia-390

Check driver installation is success

nvidia-smi

Install CUDA Toolkit 9.0
https://developer.nvidia.com/cuda-90-download-archive

sudo apt install nvidia-cuda-toolkit gcc-6 g++-6
nvcc --version

Could use update alternatives
https://askubuntu.com/questions/26498/how-to-choose-the-default-gcc-and-g-version

sudo chmod +x cuda_9.0.176_384.81_linux.run
./cuda_9.0.176_384.81_linux.run --override

Install CUDNN (ensure you are registered for the NVIDIA Developer Program)

https://developer.nvidia.com/cudnn

tar -zxvf cudnn-9.0-linux-x64-v7.1.tgz

Move the unpacked contents to your CUDA directory

sudo cp -P cuda/lib64/libcudnn* /usr/local/cuda-9.0/lib64/
sudo cp  cuda/include/cudnn.h /usr/local/cuda-9.0/include/

Give read access to all users

sudo chmod a+r /usr/local/cuda-9.0/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
sudo apt-get install libcupti-dev

Add environment variables

export PATH=/usr/local/cuda-9.0/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

Restart your terminal before proceeding to the next step. Install tensorflow-gpu

python3 -m pip install --upgrade tensorflow-gpu
python3
import tensorflow as tf
print(tf.__version__)

Reference

https://www.pyimagesearch.com/2015/07/20/install-opencv-3-0-and-python-3-4-on-ubuntu/
http://cyaninfinite.com/tutorials/installing-opencv-in-ubuntu-for-python-3/
https://linuxconfig.org/how-to-install-the-nvidia-drivers-on-ubuntu-18-04-bionic-beaver-linux