At the TensorFlow Developer Summit, Google announced that its second-generation machine learning platform has reached the v1.0 milestone. In the time since it went open-source, TensorFlow has helped many professionals in several fields make huge strides in things like language translation, early skin cancer detection, and diabetic blindness prevention.
This first major version comes with some nifty improvements. The first among those is the huge increase in speed: Accelerated Linear Algebra (XLA) will be forging the path to making models the fastest they can be. TensorFlow is now also more flexible than before, bringing with it a high-level API and includes a tf.keras module to interface with the Keras neural networks library.
Finally, one of the other foci of this revision is to make TensorFlow more production-ready by adding Python API stability (including a change to make it look more like NumPy). This will make it easier for developers to try out new features without breaking their code.
TensorFlow was just added to the latest Android Things Developer Preview, so it's really starting to expand and see some good growth. You can watch the entire eight-hour recorded livestream of the summit, or you can check out the source link for more in-depth information.
- Google Research