-
Notifications
You must be signed in to change notification settings - Fork 70
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
Add GridSearch Optimizer support #690
Conversation
…gh the grid in order
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Left a few comments. To summarize:
- let's have a separate PR with type-related fixes to the Tunable and related classes
- remove .span/.values/.quantized_values from Tunable and move that logic to the optimizer
- I am sure we can do without HashableTunableValuesDict class. I would suggest to track the tuples of tunable values in the optimizer instead of using the dict (should use less memory, too)
Also fix some missing conversion issues. Split out from #690.
Some type hint and alias tweaks for Tunables and a few others. Split off from #690.
Adds support for introspecting the values from a quantized or otherwise discrete Tunable. Split out from #690.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Left a few comments. overall, looks good, thank you!
Co-authored-by: Sergiy Matusevych <[email protected]>
Co-authored-by: Sergiy Matusevych <[email protected]>
Co-authored-by: Sergiy Matusevych <[email protected]>
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks good!
Closes #688
GridSearchOptimizer
tomlos_bench
tuple(dict.values())
fromConfigSpace
to track elements of the config grid to searchmax_iterations
>len(grid)
can refill the grid if desired (e.g., by callingsuggest()
afternot_converged()
returnsFalse
.max_iterations
<len(grid)
(i.e., we don't have enough iterations to complete the grid) will issue a warning.