TensorFlow Extended (TFX) (TensorFlow Dev Summit 2018)

Clemens Mewald and Raz Mathias present TFX, which is an end-to-end ML platform built around TensorFlow and first introduced to the public in a 2017 KDD paper. While TF.Transform and TF.Serving are already open sourced, Clemens introduces a new component, TensorFlow Model Analysis (TFMA), and give an end-to-end demo of how those tools fit together. They also announce plans about releasing more of TFX.

event: TensorFlow Dev Summit 2018; re_ty: Publish; product: TensorFlow - TensorFlow Extended; fullname: Clemens Mewald, Raz Mathias; event: TensorFlow Dev Summit 2018;

[WilsonMar]

[2:23] rather than months, "people can get up and running with it in a day and actually get to a deployable model in production in the order of weeks or in just a month"
[5:04] "training serving skew" when Jupyter notebooks correctness and functionality being dependent on their environment

[siddharthkotwal]

Really useful! We've faced pestering problems replicating results in production. How "ready" is tf.Transform to take this to production? One downside of TF seems to be that there are multiple platform, versioning errors. For example compiling the TF serving docker container for GPU is really messy, Keras + TF backend vs tf.keras DO NOT have the same performance etc.