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

Matrix-less build reduced problem in BCG #202

Open
matbesancon opened this issue Jul 2, 2021 · 0 comments
Open

Matrix-less build reduced problem in BCG #202

matbesancon opened this issue Jul 2, 2021 · 0 comments
Labels
blended_cg Blended Conditional Gradient performance

Comments

@matbesancon
Copy link
Member

We materialize a matrix M in build_reduced_problem only for an inner product x^T M y in the next function.

This matrix cannot be used in-place because its size depends on the number of atoms in the current active set.

Instead of building it, we should create a custom matrix type that:

  • lazily wraps the active set
  • implements dot(x, M, y) by looping through the active set
@matbesancon matbesancon added blended_cg Blended Conditional Gradient performance labels Jul 2, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
blended_cg Blended Conditional Gradient performance
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@matbesancon