News · March 6, 2019
TensorFlow 2.0 alpha makes Keras and eager mode the default
Google previewed a simpler TensorFlow aimed at lowering the barrier to machine learning. Teams with TensorFlow 1.x code should start planning the migration.
What happened
- The TensorFlow team announced the 2.0 alpha at the TensorFlow Developer Summit, with eager execution on by default and Keras as the core high-level API, according to its developer blog.
- The team said redundant APIs were removed, tf.function can compile Python code into graphs, and a conversion script and upgrade guide help move 1.x code to 2.0.
- 9to5Google reported that TensorFlow Lite runs on 2 billion mobile devices and that TensorFlow.js reached version 1.0 at the same event.
- The alpha is a developer preview, and the 1.x documentation remained the default during development, the TensorFlow team said.
Why it matters
The move to Keras and eager execution makes TensorFlow code easier to write, read and debug, which matters when ML work is shared across a distributed team rather than owned by one specialist.
Our view: do not port production systems to an alpha. Start new experiments in the 2.0 style, run the conversion script on a branch to see how much of your 1.x code breaks, and plan the real migration once a stable release ships. For mobile apps, keep an eye on TensorFlow Lite as on-device inference becomes practical.
Sources
- TensorFlow Blog: Test Drive TensorFlow 2.0 Alpha
- 9to5Google: Google previews TensorFlow 2.0 alpha with focus on simplicity and ML beginners
This post is CMM’s own summary and commentary on publicly reported news. Facts are attributed to the sources linked above; please read them for full details. CMM is not affiliated with these publishers or with the companies mentioned, and trademarks belong to their owners. Prepared with AI assistance. Corrections or requests: [email protected]. See our editorial policy.
Want help with this?
Tell us what you need, and we will suggest the right team model on a short call.
Talk to us