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.

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

Data Base Archive #163

Closed
pfistfl opened this issue Dec 17, 2021 · 2 comments
Closed

Data Base Archive #163

pfistfl opened this issue Dec 17, 2021 · 2 comments

Comments

@pfistfl
Copy link
Sponsor Member

pfistfl commented Dec 17, 2021

This is a feature request, based on my experiences with Optuna.

Optuna offers saving runs to a database (see here).
This allows for restarting optimization runs or re-use results from previous experiments.
(E.g. I ran a random search previously and now would like my next BO / HB / etc. to know about existing results.)

We could offer this on two levels:

  • write archive to a database, start from archive
  • write opt progress to a database

Benefits:

  • I can cancel optimization runs and do not loose all information
  • I can transfer optimization runs between machines by just porting the sqlite DB
  • I become more robust to crashes / outside influence (e.g. on servers)
  • I can at a later point just decide I want to spend more compute on something and do not loose the compute I invested before.

The API in Optuna is pretty cool, I think this might be sufficient and is simple enough and could be re-used.

@sebffischer
Copy link
Sponsor Member

@pfistfl @be-marc Maybe we could solve this with callbacks?

@be-marc
Copy link
Member

be-marc commented Jun 20, 2024

Done with ArchiveAsync.

@be-marc be-marc closed this as completed Jun 20, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants