-
Notifications
You must be signed in to change notification settings - Fork 174
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
Can't slice learning rule input? #1365
Comments
As a workaround it is possible to combine the error signals in a node and then connect that node to the learning rule. |
Though, I am not sure that the workaround actually works. I'm getting weird results, but cannot debug it right now. |
Just realized that I need this functionality more badly than I thought, so I guess I'm going to debug it after all. |
Found the problem why the workaround wasn't working: The error signal needs to be sliced in the learning rule's step function. If it is sliced in the |
This makes it easier to provide multiple error signal inputs. Fixes #1365.
#1310 added the possibility of arbitrarily sized inputs to learning rules, however it doesn't seem possible to slice into the input. I'm getting the following error:
That means I'm still stuck at not being able to implement multiple error signals (see #1307).
Or I'm doing something wrong and I'm completely blind to it ... (which is very well possible to).
The text was updated successfully, but these errors were encountered: