It's a no brainer that TensorFlow Lite is very useful for model optimization in general and thereby allowing us to make on-device machine learning as efficient and effective as possible. However, to be able to use TensorFlow Lite's latest APIs to the fullest one must be aware of many practical situations and how to tackle them. These situations include converting TensorFlow 1.x models using the latest TensorFlow Lite APIs, converting models such that they run optimally on mobile devices, and so on. In this session, we will be diving deep into practical situations like these and how to approach them.
Speaker: Sayak Paul,
Deep Learning Associate @PyImageSearch,
GDE in Machine Learning