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

Armijo rule for stochastic algorithms #1966

Open

MargaretDuff opened this issue Oct 22, 2024 · 0 comments

Assignees

Labels

Member

MargaretDuff commented Oct 22, 2024

Description

Similar to https://github.com/SyneRBI/PETRIC-SOS/blob/4e8dd56489fe22bfc832f1372c3ababeb4453bfb/main.py#L99-L156 from @samdporter we could add further options to our Armijo rule to cope in the case where the search direction is not guaranteed to be a descent direction for sufficiently small step size, such as in the case of approximate gradients in stochastic methods

For example, we could remove this error https://github.com/TomographicImaging/CIL/blob/23187c9eebfc691a4d65178c37a6cf59b01a0d57/Wrappers/Python/cil/optimisation/utilities/StepSizeMethods.py#L152C1-L155C9 so that the code still runs.

Would need some thought and maybe a check of the literature.

MargaretDuff added the enhancement label

MargaretDuff self-assigned this

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment