The Ins and Outs of TensorFlow 2.0 - PowerPoint PPT Presentation

About This Presentation
Title:

The Ins and Outs of TensorFlow 2.0

Description:

Make the most of this latest iteration by partnering with seasoned TensorFlow developers. Source Url:- – PowerPoint PPT presentation

Number of Views:26

less

Transcript and Presenter's Notes

Title: The Ins and Outs of TensorFlow 2.0


1
The Ins and Outs of TensorFlow 2.0 
2
  • In January 2019, Google officially announced
    that the new iteration of TensorFlow will be
    available in September 2019. Since the
    announcement of this news, there were umpteen
    questions popping on the minds of TensorFlow
    developers. Some were wondering that the latest
    iteration TensorFlow 2.0 would have a large jump
    similar to AngularJS v2 vs AngularJS v1. While
    others were thinking that the graphs would work
    properly in TensorFlow 2.0. 
  • In this informative piece, we will walk through
    the changes that are worth noticing in TensorFlow
    2.0. Lets get started. 

3
Noticing in TensorFlow 2.0
  • Clean API
  • Complete Control Over Variables
  • Graph Mode Functions

4
Clean API
  • TensorFlow 2.0 has been redesigned and in the new
    iteration many APIs have shifted to separate
    repositories or removed including tf.app,
    tf.gans, tf.flags, tf.contrib, and tf.logging.
    While some APIs of TensorFlow 1.0 have been
    replaced with their TensorFlow 2.0 counterparts
    such as tf.keras.optimizers, tf.keras.metrics,
    and tf.summary. In TensorFlow 2.0, tf.keras is an
    advanced API that is strongly recommended. 

5
Complete Control Over Variables
  • In TensorFlow 1.0, data scientists used to keep a
    tab on the variables for future use. In other
    words, the older version banks heavily upon
    implicit global namespaces. So, if you were not
    present during the initial development stage, you
    would struggle to recover something that you
    never knew existed. However, TensorFlow 2.0 has
    enabled you to overcome all such hassles. 
  •  

6
Graph Mode Functions
  • The tf.function() in TensorFlow 2.0 enables data
    scientists to execute functions as a single
    graph. This operation allows TF 2.0 to leverage
    the benefits of graph mode such as optimized
    functions for kernel fusion or node pruning,
    resulting in enhanced portability of functions -
    import or export. 
  •  

7
Wrapping Up
  • This informative piece walks you through the ins
    and outs of the latest iteration TensorFlow 2.0.
    It comes with many advanced features with a focus
    on simplicity, developer productivity, and ease
    of use. Though in this piece, we discussed only a
    few of them. Hope you find them useful. If you
    want to make the most of TensorFlow 2.0, its
    wise to hire TensorFlow developers from a trusted
    partner. 

8
Source Url-
https//damcotechservices.wordpress.com/2021/02/11
/the-ins-and-outs-of-tensorflow-2-0/
Write a Comment
User Comments (0)
About PowerShow.com