SentencePiece is the first step to many modern state of the art Natural Language Processing models like BERT. We will understand how this component works, (reading the paper beforehand is advised: https://arxiv.org/abs/1808.06226), then use some tools to use it and see how it affects our data and finally deploy it to Kubernetes using TensorFlow-serving.

Requirements: General Knowledge of NLP process, TensorFlow, examples will be in Python

About Presenter: Fabrizio Milo is a passionate deep learning practitioner, located in Los Angeles, he actively participates to various meetups on deep learning and TensorFlow. He adopted TensorFlow from the first releases and contributes to the project regularly. He is interested in distributed computing and acceleration. Fabrizio holds a Master in Computer Science from the University of Rome “La Sapienza.”