-
Notifications
You must be signed in to change notification settings - Fork 331
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Minor changes #501
Minor changes #501
Conversation
Thanks a lot. |
Thank you, @SimonInParis
|
@mahbodnr, Thank you for your help I agree, using STDP with the default value needs to raise a warning to the user that About Please fix the issues above and I will accept the PR. |
@Hananel-Hazan Sure, thanks. |
@Hananel-Hazan done. |
Thanks @mahbodnr for your help |
A few suggestions:
Nodes:
Whentraces_additive
is False, spike trace (self.x) would be filled with 1 at spike points. However, it is not always the case (For example in Izhikevich 2007, this value is set to 0.1). I think the 1 is better to be replaced by thetrace_scale
value, so users can set this value as well.Nodes:
typo:self.tau
Learning:
the default value fornu
is set toNone
, which makes thePostPre
learning rule practically wasteful. I believe it is not a good default value for an argument and it is better to be replaced by some float numbers or get changed to a mandatory argument.Network:
the format of theinputs
argument of therun
method is not very common and since the input in most of the deep learning frameworks (mainly TensorFlow and Pytorch) is in the type of tensors, it can be tricky especially for new users. Furthermore, since there is no explicit error to point out this problem, it could be sometimes hard to find and a time-consuming error. Adding an assert to check the type ofinputs
can be helpful in this regard.Network:
A progress bar can be useful in very long runs A custom progress bar (including some information about the run and the network) can be added later.Learning:
replacetorch.ger
withtorch.outer
(https://pytorch.org/docs/stable/generated/torch.ger.html)models:
typo:inhibitory
@Hananel-Hazan