Today Deep Learning holds state-of-the-art performances in many tasks, but mainly supervised training with huge and fixed datasets. Continual Learning aims to give machine learning models the ability to learn continuously whenever a new task is assigned considering the past experiences.