-
Notifications
You must be signed in to change notification settings - Fork 78
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
Unify duplicate implementation of CAN #158
Comments
the following note was left on the notebook Notes:The attention function provided for us does not normalize the attention coefficients.Should this be done?Where should we be able to customize the non-linearities?Seems important for the output.What about the attention non-linearities do we just use what is given?I wanted to make this so that without attention it ends up beingthe Hodge Laplacian network.Maybe ask the contest organizers about this?I agree on the above. Attention in the base layer is used in different context typically used in the lit. lit : it means learnable parameter that is multiplied by the original vector during training. Sum = 1 for a given neighborhood(x) In general the layer is not faithful to the CAN implementation so I recommend deleting it from the package. |
Sounds good. The layer that is not faithful to the CAN implementation can be deleted. Can you open a PR that does it? Once that PR is merged, and the corresponding CAN code is removed, we can mark this task as done. |
What?
There were two implementations of CAN by challenge participants.
If they are indeed the same implementation, we need to merge them into 3 files:
If they are different implementations, we need to check if one of them can be deleted.
Possible reasons to delete:
If we decide to keep both of them, the file names and docstrings should be updated to highlight their difference.
Why?
Because it is confusing to have two implementations of CAN without knowing their difference.
Where?
The files to modify are:
How?
See for example how the duplicate SCN and SCN2 has been solved and the use of the "See Also" sections in their docstrings.
Additionally:
_v2
incan_layer.py
because:The text was updated successfully, but these errors were encountered: