Regarding the model file: -> Currently, the prediction model will be copied to the image file system if it exists; otherwise, the `create_model.py` script will create it directly in the image if it did not exist. -> This would not be acceptable in the case of a model which has to be trained, but this could be bypassed by storing the model in a cloud location, for the script to download (which was not the case in the present task) Starting the application: -> docker-compose build [--build-arg LOG_LEVEL=VALUE] classification -> `LOG_LEVEL` must have one of the following values: 'DEBUG', 'INFO', 'WARNING', 'ERROR' or 'CRITICAL' -> If the build-arg option is not specified or the value provided for `LOG_LEVEL` is not one of those described, the default level of 'DEBUG' will be used -> docker-compose up classification The web application is now running on http://localhost:8060 Calling the web-service: -> From the web browser, visit http://localhost:8060/swagger for viewing the Swagger UI and testing the endpoint -> From the terminal, use `curl -F image=@fileName http://localhost:8060/classifyImage` for classifying the specified file using the VGG 16 classifier