Skip to content
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

Deep R #94

Draft
wants to merge 21 commits into
base: master
Choose a base branch
from
Draft

Deep R #94

wants to merge 21 commits into from

Conversation

neworderofjamie
Copy link
Contributor

No description provided.

Base automatically changed from genn_5 to master April 25, 2024 14:20
@codecov-commenter
Copy link

codecov-commenter commented Aug 30, 2024

Codecov Report

Attention: Patch coverage is 28.57143% with 100 lines in your changes missing coverage. Please review.

Project coverage is 64.00%. Comparing base (af2d5da) to head (62e316e).

Files with missing lines Patch % Lines
ml_genn/ml_genn/compilers/eprop_compiler.py 9.37% 29 Missing ⚠️
ml_genn/ml_genn/compilers/deep_r.py 40.47% 25 Missing ⚠️
ml_genn/ml_genn/compilers/event_prop_compiler.py 9.09% 20 Missing ⚠️
ml_genn/ml_genn/callbacks/custom_update.py 45.00% 11 Missing ⚠️
ml_genn/ml_genn/utils/model.py 27.27% 8 Missing ⚠️
ml_genn/ml_genn/compilers/compiler.py 41.66% 7 Missing ⚠️
Additional details and impacted files
@@            Coverage Diff             @@
##           master      #94      +/-   ##
==========================================
- Coverage   64.80%   64.00%   -0.81%     
==========================================
  Files         101      102       +1     
  Lines        4259     4342      +83     
==========================================
+ Hits         2760     2779      +19     
- Misses       1499     1563      +64     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

* When building optimiser, can pass through EGP references to track sign changes in either direction
* Added function to build Deep-R custom connectivity updates
* E-Prop compiler takes list of Deep-R excitatory and inhibitory connections
* New ``CustomUpdateOnTrainBegin`` callback to launch an initial custom update
* DeepRInit should trigger at beginning of first epoch, not at start or training (in fact, start of training is a silly time to start any kind of custom update)
* Refactor callbacks to take a filter function and replaced bespoke callbacks used in EventProp compiler
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants