Starter kit for assignment 3 of the PCDTV course. The goal of this assignment is to train from scratch a small neural network to extract speaker embeddings. The learnt embeddings are then evaluated for the task of speaker verification. For more information, please see the associated Colab notebook.