Mistake in AdaDelta?

[chrismaliszewski]

I'm implementing AdaDelta. I think there is a mistake in your function.

I found a difference between your code and the official implementation:

I also looked at other implementations:
https://d2l.ai/chapter_optimization/adadelta.html
https://gluon.mxnet.io/chapter06_optimization/adadelta-scratch.html

You first calculate both E_g and E_p and based on them the delta whereas you should just calculate E_g, then delta and finally E_p.

Am I correct that there is a mistake in your code?