Conversation
E.g. data_reader, config_package, input_list, conns_to_include
So the optimizer can also add new parameters instead of only changing the magnitude. Therefore the unit needs to be known.
parameters['AVBR_to_DB1_elec_syn_gbase'] = {'value':'0.1', 'unit':'nS'}
self.volts['MDL01/0/GenericMuscleCell/v'] = [...]
Check if type of param is dict or list (new or old version).
Use multiprocessing.pool instead of parallelpython
This reverts commit 87bf234.
Conflicts: CElegans/pythonScripts/c302/tune/C302Simulation.py
Only override/add param by a regex unless there is a param defined without a regex.
new models, other changes
| vm2pR , VC03 ,2,electrical | ||
| vm2pR , um1pR ,4,electrical | ||
| vm2pR , vm1pR ,4,electrical | ||
| vm2pR , vm2pL ,4,electrical No newline at end of file |
There was a problem hiding this comment.
@lungd Why are these changes being added to herm_full_edgelist.csv?
Surely this should remain identical to https://github.com/openworm/PyOpenWorm/blob/dev/OpenWormData/aux_data/herm_full_edgelist.csv (or the reason for the change documented...)
| min_constraints = [] | ||
| max_constraints = [] | ||
|
|
||
| parameters['AVBL_to_AVBR_exc_syn_conductance'] = {'default_unit': 'nS'} |
There was a problem hiding this comment.
@lungd Are these necessary here? Fine to keep them as long as they're not temp additions... Have you checked to see if some of the other optimisation examples in these (tuning) files still work?
|
@lungd I've manually added all these changes on the master branch, apart from the changes to the tuner code (which I still need to retest on mine; but it would be good if you double check if these are necessary/correct in your changes) and the additions to herm_full_edgelist.csv (see above). Can you merge these into yours and test? They'll require the latest sibernetic_c302 (here). Btw, all of my updates are working in docker in this branch: https://github.com/openworm/OpenWorm/tree/development_c302 (if you want to try them out too @slarson) |
3 participants
Hi guys,
For the version of the OpenWorm dockerfile, I'd like to incorporate David's changes. This will allow us to have the dockerfile point at the CElegansNeuroML repo and therefore centralize some of the many versions that have been floating around. Please review and let's merge ASAP.
Thanks,
Stephen