app.py contains the Flask framework that parses GET/POST information, keeps track of the image annotations, trains the Machine Learning model and runs the predictions.
ml_model.py contains the Python Machine Learning code that takes the pretrained neural net and retrains the last layers with the annotated Tufa examples. This file is based on code from the Tensorflow github.
ml_model contains the pre-trained models cached from https://tfhub.dev/google/imagenet/mobilenet_v2_100_224/feature_vector/2 and https://tfhub.dev/google/imagenet/mobilenet_v2_035_128/feature_vector/2
static contains the image JPG files.