Skip to content
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.

Sign up for GitHub

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

Jump to bottom

Implementation details #2

Open
clementabary opened this issue Mar 9, 2020 · 0 comments
Open

Implementation details #2

clementabary opened this issue Mar 9, 2020 · 0 comments

Comments

@clementabary
Copy link

Thanks a lot for the great work.

I still have a few questions on implementation details.
First, what is the reason for partitioning the training procedure with powers of 2 ?

Second, I am confused with normalization. For the source dataset, you use the maximum absolute value normalization while for the sample dataset you use a scaling with lambda x: 10 * 1.0 / x.pow(2).sum().sqrt(). Can you give more insight on this choice ?

Third, why did you choose to pad your sequences with small noise rather than zeros ?
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@clementabary