TensorFlow 1.14.0 has been released!

[Sundeep Gottipati]

TensorFlow 1.14.0 has been released! Please see the full release notes for details on added features and changes.

Here is a brief summary of the major improvements in this release:

• This is the first 1.x release containing the compat.v2 module. This module is required to allow libraries to publish code which works in both 1.x and 2.x. After this release, no backwards incompatible changes are allowed in the 2.0 Python API.

• Turn on MKL-DNN contraction kernels by default. MKL-DNN dynamically dispatches the best kernel implementation based on CPU vector architecture. To disable them, build with --define=tensorflow_mkldnn_contraction_kernel=0.

• Non-Windows system libraries are now versioned. This should be a no-op for most users as it affects only system package maintainers or those building extensions to TensorFlow.

Thanks!

[Austin Anderson]

Hey Harald,

Sorry about the delay. The images are indeed available now (as of this morning), and were missing due to some confusion in our CI system due to the coincident release of 2.0b1. You can expect to see more timely image releases for future TensorFlow versions, as I'll be fixing that particular bug.

Best,

Austin

On Thu, Jun 20, 2019 at 12:02 PM Sundeep Gottipati <gsun...@google.com> wrote:

Actually +Austin

On Thu, Jun 20, 2019 at 12:01 PM Sundeep Gottipati <gsun...@google.com> wrote:

Sorry about that, it rebuilds them from the latest pip tag which is now determined to be "2.0.0b1".

+Austin, can you manually build these 1.14 images? We should be building images for any pip tag without a docker release - not just the latest tag.

On Thu, Jun 20, 2019 at 5:43 AM <harald...@gmail.com> wrote:

Can't seem to find any tf docker images tagged 1.14. When can we expect them?