search

dekamps / miind

http://miind.sf.net
MIT License
13 stars 7 forks source link

Blackboard #20

Open MartinPerez opened 7 years ago

MartinPerez commented 7 years ago

Here are my cleaned modifications, I am missing documentation and for activity circuit to read parameters from file. Baseline circuit and partial activity circuit are just copies of Activity Circuit, since I was not reading parameters from a file. That should go away.

dekamps commented 7 years ago

Hi Martin,

I messed a lot with the master version, but the tar ball should be fine for your purposes. Nonetheless, I will dive into the code quickly 😉

From: Martin Perez-Guevara notifications@github.com Sent: 06 June 2017 13:22:26 To: dekamps/miind Cc: Subscribed Subject: [dekamps/miind] Blackboard (#20)

Here are my cleaned modifications, I am missing documentation and for activity circuit to read parameters from file. Baseline circuit and partial activity circuit are just copies of Activity Circuit, since I was not reading parameters from a file. That should go away.

You can view, comment on, or merge this pull request online at:

https://github.com/dekamps/miind/pull/20

Commit Summary

File Changes

Patch Links:

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/dekamps/miind/pull/20, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AEk-Y8FTDiTW6-4_rBdPAigLiOvKuzwCks5sBUSCgaJpZM4NxRAv.

MartinPerez commented 7 years ago

Hi @dekamps, I just finished adding documentation of the main script activity_circuit.cpp. It should be almost trivial to pass the whole simulation to adaptive neurons, since only the class EIB_Population_Pair needs to be modified with the appropriate rate update algorithms :).

On the other hand I had some issues passing an algorithm class as argument to objects, which is why the algorithm definition is inside EIB_Population_Pair, even better would be to simply pass the desired algorithms as argument to the classes, such that they just implement the architecture concept. I should also take outside of the BBCell class the node information added to a csv file and just create a method to do this. It might also be good after some refinements to bring all the blackboard classes to their own file, separated from the simulations that are implemented in main. All these are details for the code re-usability, nothing crucial. Please let me know what you think.

PS: Locally I can run without problems in the older version. This will just be necessary to merge the Blackboard example into master. It would be nice to pass all the recent changes to a development branch and leave master only for stable updates, like a pull request passing tests and compilation.