• Add opencv_contrib to the Dockerfile.
• Revise cmake commands and add flags.
• Add Libtorch to the Dockerfile.
• Add sample test codes to test the libraries installed. Assert that they should pass.

See: krshrimali/DCGAN-CelebA-PyTorch-CPP#15

Summary: Attempt to avoid building OpenCV from source, and using libopencv-dev instead.

TODOs:

• Make a PR applying these changes. cc: @krshrimali
• Test these changes (once 1st is done, i.e. PR is made) by building docker and running the codes for DCGAN project. cc: @HimanshuSinghGH

It will be great if the steps can be added to run the Dockerfile (in the README file)

cc: @HimanshuSinghGH

The spelling of container is wrong. "Dockerfile for a 'contained' installed" should be "Dockerfile for a 'container' installed".

Just like we have tests for checking PyTorch and OpenCV imports in Python (see below), it will be great to have tests for these libraries in C++.

python3 -c "\
try:
    import cv2
    print('OpenCV was installed successfully')
except ImportError:
    print('Error occured, file issue on GitHub repo. OpenCV not installed successfully')"

python3 -c "\
try:
    import torch
    print('Torch was installed successfully')
except ImportError:
    print('Error occured, file issue on GitHub repo. PyTorch installed successfully')"

• Add tests for PyTorch in C++
• Add tests for OpenCV in C++

Update: No need to add tests for Libtorch since we explicitly link to CMake config files using CMAKE_PREFIX_PATH flag.