This project addresses the science and engineering of developing few shot learning models, with the aim of reducing the data required to train machine learning models and the size of the models.