Analysing the output data (supercomputers)

A Python package that facilitates development and use of models of biological neural networks

Moderator: tom_morse

Post Reply
parvinz
Posts: 1
Joined: Mon Oct 22, 2018 9:16 pm

Analysing the output data (supercomputers)

Post by parvinz » Wed Oct 24, 2018 11:03 pm

Hi,

I am using supercomputers to run the model that I created using NetPyNe.
I have a problem with analysing the output data.
I tried using the tut8_analysis.py file mentioned in the tutorial, but it seems that one of the functions are missing (df.popRates).
I was wondering how I could plot the same figures using 'plotConn' and 'plotTraces' on the output of supercomputers.

Many thanks.

salvadord
Posts: 76
Joined: Tue Aug 18, 2015 3:49 pm

Re: Analysing the output data (supercomputers)

Post by salvadord » Thu Nov 22, 2018 1:35 pm

Hi,

The tut8_analysis.py was just provided as an example of how to analyze output data. It may not work with all models. The reason that df.popRates might be missing is that you need to set cfg.printPopAvgRates = True -- this records the avg population rates and stores them in simData.

If you want to get the 'plotConn' and 'plotTraces' from the output of the supercomputer, you can load the output files using sim.load(filename). However, this assumes that the output file contains all the required data, including list of cells and connections. Once the file is loaded, you can use sim.analysis.plotConn() and .plotTraces() to get the plots.

Alternatively, you can generate these plots directly from the supercomputer simulations. Just add these plots to cfg.analysis (as in the tutorials) and set the options saveFig=True, showFig=False.

Let me know if you have further issues.

Post Reply