This is 
In tf.keras in TF 2.0, how do you write new layers with complicated behavior? Say, a BatchNormalization layer? Answer: very simply and very intuitively. Basically the way you'd write Numpy code. No more symbolic control flow. https://colab.research.google.com/drive/1P4ejqu5W7iKsEiSoPwNd5bQ5fVbKi2rq …pic.twitter.com/p8HoVhBgvg
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Is this under eager_execution? :o how is the performance compared to the "classic" way?
-
This work with both eager and graph execution. The Notebook above uses graph execution by default.
End of conversation
New conversation -
-
-
Basically the way you'd write PyTorch code, to be precise
-
Every time you wrote pytorch and it has no sense die a cat. If you did it just to troll
@fchollet you will need to try and better
- Show replies
New conversation -
-
-
Are you going to write a new book with the new updates ?
-
Yes, 2nd edition of Deep Learning with Python
- Show replies
New conversation -
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
That is god damn beautiful. Thank you for making neural networks easy and approachable
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What's the point? Pytorch already does all that.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.