Differences Between Tensorflow 1.x and Tensorflow 2.0 (Episode 3)

*** Get the limited-time VIP version of the course HERE ***

[vil]

I wanted to know what is eager execution and landed in this video. Found the explanation simple and crystal clear. Wow! Thank you!

[xdxn]

Thanks, great talk. Why @tf.function and Session() paradigm more efficient than eager excution?

[moseswai-mingwong]

I learnt TF 1.13 few years ago, and hated the wried session computation graph construction model, Google know it and while Keras takes off since most of us are happy to use pre-build modules from Keras, here comes TF 2. 0 which supports all the state-of-the-art requirements - simplicity and beauty of Keras, eager execution. Now AI code written in TF 2.0 is just cleaner and better. Thanks for this superb video, in just 10 minutes you cover everything! Great job lazy programmer!

[viinisaari]

Loving the shift to eager execution! The whole session paradigm was always my biggest complaint about tensorflow

[smrgardening]

as you said....Keras is now official API, then what you suggest?? should i start using Keras and ignore tensor flow??
actually i already converted my all codes into Keras and i am in love with it's intuit.
hope you'll also convert all your tutorials into Keras and from now on make Keras tutorials.

[bzqp]

It always confused me - how did TF1 know, that when I was using the "+" command I didn't mean the python "+" but the tf "+"? How did TF know I'm not just adding an integer to a tensor, but I'm instead making a computation graph that adds the integer?

[gesuchter]

I've got a book on TF1 which I haven't read yet. Would you recommend buying a book with TF2 or is it okay to start with TF1 and later migrate to TF2?

[mohammedinnat]

You're a lazy reader but a great explainer. Thank you. :D