concept variety in category deep learning

appears as: variety
Deep Learning with Structured Data

This is an excerpt from Manning's book Deep Learning with Structured Data.

In this book we’re going to use Keras as the deep learning library in the stack summarized below. Keras began life as a self-standing project that could be used as a front-end for a variety of deep learning frameworks. As explained in chapter 5, as of TensorFlow 2.0, Keras is integrated in TensorFlow. Keras is the recommended high-level API for TensorFlow. The code accompanying this book has been validated with TensorFlow 2.0 but you should not have any issues using later versions of TensorFlow.

Keras began its life as a front-end for a variety of backend deep learning frameworks, including TensorFlow and Theano.  The purpose of Keras was to provide a set of accessible, easy-to-use APIs that developers, particularly those who were learning deep learning for the first time, could use to explore deep learning. When Keras was first released in 2015, the deep learning backend libraries that it supported (first Theano and then TensorFlow) provided a broad range of function but could be challenging for beginners to master. With Keras, developers could get started with deep learning using familiar syntax and without having to worry about all the details exposed in the backend libraries. If you were starting a deep learning project back in 2017, your choices included: