-
Notifications
You must be signed in to change notification settings - Fork 357
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Register only spikes actually sent with weight_recorder #3044
Conversation
…on<T>::send() must return whether they sent a spike or not.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I applied these changes to a copy of the latest eprop_feature
and applied the corresponding changes to eprop_synapse.h
(returning false if the spike is to be ignored) and learning_signal_connection.h
.
All verifications were successful.
By themselves, and all in all, the changes seem very appropriate
Co-authored-by: Dennis Terhorst <terhorstd@users.noreply.github.com>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM 👍
Until now, spikes would be registered with a
weight_recorder
even if the spike was not transmitted to the target, e.g., by abernoulli_synapse
. This PR changes allConnection<T>::send()
methods to returntrue
if they actually transmit a spike to the target node (calle();
) andfalse
otherwise.The PR also simplifies the code in the
bernoulli_synapse
. Since this synapse model can only be used between normal nodes (not to/from devices), it will only receive events with multiplicity 1, allowing for significant simplification.@JesusEV @terhorstd I'd appreciate if you could fast-track this PR since e-prop work depends on it.