Abstract

Creating a Convolutional Neural Network can be very complex, especially for people who have no programming experience.
A tool to generate the project code makes it more likely for non-experts to give Deep Learning a try. Having a Configuration
Language enables the users to easily create their model and tweak the parameters in a plug-and-play environment using puzzle
pieces. These different parameters determine the viability of your project, so thoughtfully setting these is advised. A user will
often have a dataset that is too small to generate useful models. Therefore, Data Augmentation code can be provided by the tool
if parameters concerning the data are given. This paper discusses a possible solution.
Demo Video

  Reading Phase

Reading Report (PDF)
Reading Presentation (PDF)

  Implementation Phase

Final Report (PDF)
Final Presentation (PDF)
Project Materials (ZIP)