Fabian Pedregosa

Thanks for the report. For the timing, having numba installed can help. In your notebook you can install it with "!pip install numpy scipy numba" . However, I can see...

@GeoffNN : any idea why it's so slow? Could it be related to the use of sparse matrices?

I'm building the doc locally to see. I'm traveling at the moment so it might take a while. Hopefully by tomorrow we'll know On Tue, Jan 24, 2023, 17:48 busycalibrating...

out of curiosity @busycalibrating , have you tried running the example locally? It could also be that colab's free tier has caps CPU speed or other resources

All yours @alirezadamash !

All yours @mmhamdy !

hey @mmhamdy , any progress on this? I might have some cycles to work on this issue next week if it's not yet solved

Excellent, and thanks for the update! I'm trying to push for this as I'd like to test on Python 3.12 but unfortunately TF is blocking us from doing so 😞...

> Also, a small note, mlp_mnist.ipynb and differentially_private_sgd.ipynb appear on the gallery but don't appear on the side menu. Nice catch, I've opened https://github.com/google-deepmind/optax/issues/767 to track this

thanks a lot @innat for this contribution! It would be indeed great to have an example notebook for those that wish to use keras. However, for the competition, participants will...