Closed russelljjarvis closed 1 year ago
the only code that needs to be in this PR is that in src/contrib/genPotjansConnectivity.jl. and that should be split into two files, one with the definition of genStaticWeights and the other with genPlasticWeights. the rest is just your test harness and shouldn't be committed.
AOK, that makes sense, I will revise and update.
It was tempting to try to put both the genStaticWeights
and genPlasticWeights
in the same file for computational tractability reasons, as they both could be made in the one pass of doubly nested for loop, and lumping them together would reduce the run time complexity of the code.
From a different perspective as of this moment I am not sure if the Potjan's model has plastic weights, and I think the default genPlasticWeights
erdos method would be just as appropriate, as an approach that just finds idle static synapses.
I still remember this PR. I am currently pre-occupied with lab work, but I am also incentivized by my lab to make this PR suitable.
Okay. I have done rebasing and boiled this down to six commits.
I have also synced back up with your branch.
Abandoned in favor of new less commits alternative. https://github.com/SpikingNetwork/TrainSpikingNet.jl/pull/7
Another draft PR.
This code shows how Plugin matrix capabilities almost work.