Over the last few years, Tensorflow Team added a number of components to the Tensorflow.
With TensorFlow 2.0, these will be packaged together into a comprehensive platform that supports machine learning workflows from training through deployment.
Install the TensorFlow 2.0 Beta preview package:
pip install tensorflow==2.0.0-beta0
TensorFlow 2.0 will focus on simplicity and ease of use, featuring updates like:
- Easy model building with Keras and eager execution.
- Robust model deployment in production on any platform.
- Powerful experimentation for research.
- Simplifying the API by cleaning up deprecated APIs and reducing duplication.
Let’s take a look at the new architecture of TensorFlow 2.0 using a simplified, conceptual diagram as shown below:
Tensorflow 2.0 beta
TensorFlow 2.0 Beta is designed to let developers, enterprises, and researchers easily build and deploy ML powered applications with an emphasis on usability. It focuses on model building using Keras as the high-level API, eager execution as the default for intuitive development and debugging, and @tf.function for graph-like performance and portability.
What's new in beta?
In this beta release you'll find a final API surface, also available as part of the v2 compatibility module inside the TensorFlow 1.14 release. We have also added 2.0 support for Keras features like model subclassing, simplified the API for custom training loops, added distribution strategy support for most kinds of hardware, and lots more. You can see a list of all symbol changes here, and check out the link below for a collection of tutorials and getting started guides.
Thank You !