Conversation
bindsnet/pipeline/__init__.py
Outdated
| else: | ||
| self.history = {} | ||
|
|
||
| # monitor spikes for selecting action based on sum of spikes |
There was a problem hiding this comment.
It seems like this shouldn't be done in the Pipeline object, but outside in an experimental script. If the layers of the network are really large, for example, this could make things really slow.
There was a problem hiding this comment.
Yes. You are right. I will make the change.
| @@ -0,0 +1,56 @@ | |||
| import torch | |||
There was a problem hiding this comment.
What's the purpose of this script? We want to have only a few examples in the BindsNET repo so we don't have to maintain a bunch.
There was a problem hiding this comment.
@Hananel-Hazan wanted an experiment for reinforcing only the right action via STDP. This was to see if it performs well enough to include the results in the paper. At this point, I am not really sure which one we should use. The breakout.py example performs the best. Which is not saying much because it has no learning.
There was a problem hiding this comment.
Got it, thanks. It might be better to keep this script in the bindsnet_experiments repo.
There was a problem hiding this comment.
Ok. I will remove it for now. Since I am not sure about the utility of this experiment.
There was a problem hiding this comment.
It should be in directory with the rest of the benchmark scripts and example scripts that will appear in the paper for Frontier. @djsaunde will create a spacial tag for it
3 participants
Major changes: