PyVertical Development of SplitNN as a 2d array of Models.

Development of SplitNN as a 2d array of Models.

Open H4LL opened this issue 3 years ago • 6 comments

This allows for generic logic which can be used to accomodate arbitrary input/ label tensor partitions.

Currently we treat the SplitNN as a 1d array of models. This allows us to perform horizontal splits in the model.

When this works, we should be able to use the same, standard class for any data/ label distribution

Aug 18 '20 13:08 H4LL

Hi @H4LL Can I work on this issue, please?

Aug 26 '20 06:08 arunraja-hub

Sure! Lmk know if you'd like any pointers/ feedback

Aug 27 '20 16:08 H4LL

Thank you! This is my first issue on Openmined Could you point me to some articles/resources I can refer to learn more about this SplitNN project?

Aug 27 '20 16:08 arunraja-hub

https://blog.openmined.org/split-neural-networks-on-pysyft/ https://blog.openmined.org/what-is-pyvertical/

Aug 28 '20 13:08 H4LL

hey @arunraja-hub, how is this going?

Feb 10 '21 15:02 TTitcombe

Apologies I haven't had time to work on this You can reassign it to someone else

Feb 11 '21 13:02 arunraja-hub