Performance/speed improvements between versions

ztlkl.cz · July 23, 2021, 8:41am

Hi, currently we are using Python 3.6 by Enthought with TensorFlow 2.2.2. We are going to update to Python 3.8 with TensorFlow 2.5+.

Are there any improvements (speed/performance) beteen these version of TensorFlow or Python?
I can’t find any relative information about this topic so maybe you can just point me on right place.

Bhack · August 4, 2021, 10:44pm

Some time ago there was a proposal at:

github.com/tensorflow/build

How easy it is to externalize the visibility of the benchmarks we already run continuously?

opened 05:23PM - 02 Jul 20 UTC

bhack

This a pinpoint to not lost the benchmarking subthread started at https://github….com/tensorflow/tensorflow/pull/33945#issuecomment-652471742 Quoting @zongweiz > We have [TF2 benchmarks] (https://github.com/tensorflow/models/tree/master/official/benchmark) from every official model garden model and they continuously running internally, using PerfZero framework. We should be able to select a set of benchmarks and hook it up with github CI (jenkins). Added @sganeshb to this thread for information and we will follow up. Another thought is to start CI tests using a selected set of tf_benchmarks /cc @alextp @naveenjha123 @sganeshb

As you can see this is a topic that could interest also more in general the community on some specific type of contribution:

/cc @markdaoust I don’t know if you have a specific team member to mention here as I don’t have a one-o-one mapping between our github nicks involved in a thread and the ones on the forums