TensorFlow Addons is a collection of extra functionalities gathered into a single repository (https://github.com/tensorflow/addons). It hosts some of the newer advancements in deep learning that are too unstable or not used by enough people to justify adding them to the main TensorFlow library. It also acts as a replacement for tf.contrib, which was removed from TensorFlow 1.
TensorFlow Extended is an end-to-end machine learning platform for TensorFlow. It offers several useful tools:
- TensorFlow Data Validation: A library for exploring and validating machine learning data. You can use it before even building your model.
- TensorFlow Transform: A library for preprocessing data. It allows you to make sure training and evaluation data are processed the same way.
- TensorFlow Model Analysis: A library for evaluating TensorFlow models.
- TensorFlow Serving: A serving system for machine learning models. Serving is the process of delivering predictions from...
