All the paths in .py files are changed to 3 new directories.
Instructions of running the docker container are inserted into the README.
I successfully built the docker image(training the model on my own pictures) on my windows laptop, but I could not run a container from it(testing the program) as windows does not give camera access to docker containers. Then, I tried to build the image based on linux/arm64/v8 platform to test it on my RaspberryPie device, but it could not install Tensor Flow V1.7 in this setting.