ANN Challenge team JAN

I rearranged the repository structure, added a .gitignore file and split the code into two scripts:

• script.py trains the network with the frozen VGG16 model it is called like this: python3 script.py model_name.hb5 where model_name.hb5 is the file where the script will save the first model
• fine_tuning.py is used for tuning the previous model. It is called like python3 script.py model_name.hb5 where model_name.hb5 is the name of the model it will fine-tune. The final model is saved as model.hb5