Left, a GAN training loop based on the GradientTape, and right, a GAN training loop based on train_on_batch (from my 2017 book).pic.twitter.com/RcSboLBw6O
Show this thread
The left version is a bit more verbose, but that's because it's significantly lower level. The right version would require a bit more background to understand.
-
-
Do you have any plan to update your book or should I buy the current version?
-
There will be a second edition, presumably released at the end of 2020.
- Show replies
New conversation -
-
-
It would be nice to see the same comparison with a GAN using gradient penalty. That seems easiest by far to implement using tf.gradients which won't work in eager mode, but maybe there's a nice way to do it with GradientTape too.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Helpful
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.