If you create something cool with TF 2.0, tweet it to me, I'm always interested! Likewise, if you have any feedback about tf.keras or TF 2.0 in general, please let us know.
Show this thread
-
I'm still beginner on Deep Learning. Sorry to ask silly question Currently, i tried to moving my RL project to TF 2 from pytorch. but i'm still confused about tf.function. Must we use that thing on the function with tf.GradientTape() inside it or only use it on loss function ??
Think of tf.function as a way to compile code into a graph. It will run faster, but creating and dispatching it has some fixed overhead. So you should create as few tf.functions as possible, usually just one wrapping your entire training step, like this:https://twitter.com/fchollet/status/1105141298825093122?s=19 …
-
-
Thank you for your answer. It's very clear and understable
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.