Automated Setup Script for CUDA, cuDNN, and OpenCV Installation

This script automates the installation process for CUDA, cuDNN, and OpenCV on an Ubuntu system. It provides a streamlined way to ensure that the required dependencies and libraries are properly installed for deep learning and computer vision development.

Requirements

Ubuntu operating system (tested on Ubuntu 20.04)
NVIDIA GPU with CUDA support

Usage

Clone this repository to your local machine:

git clone https://github.com/mhkarazeybek/install_cuda_cudnn_opencv.git

Navigate to the project directory:
```
cd cuda-cudnn-opencv-installation
```
Make the script executable:
```
chmod +x install_script.sh
```
Run the script with necessary arguments to specify the versions:

Replace <CMAKE_VERSION>, <CUDA_VERSION>, <CUDNN_VERSION>, <OPENCV_VERSION>, and <UBUNTU_VERSION> with the desired versions of CMake, CUDA, cuDNN, OpenCV, and Ubuntu respectively.
```
./install_script.sh --cmake <CMAKE_VERSION> --cuda <CUDA_VERSION> --cudnn <CUDNN_VERSION> --opencv <OPENCV_VERSION> --ubuntu <UBUNTU_VERSION>
```
Follow the on-screen instructions and prompts.

Notes

The script first uninstalls existing CUDA, cuDNN, and NVIDIA drivers if they are detected.
It then installs the necessary NVIDIA drivers automatically using the ubuntu-drivers command.
The script compares the installed CMake version with the provided version. If the installed version is older, it removes the old version and installs the specified CMake version.
CUDA, cuDNN, and OpenCV are installed based on the provided version numbers.
After successful installation, the script verifies the OpenCV installation.