-
Notifications
You must be signed in to change notification settings - Fork 2
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
Grid locator #20
Grid locator #20
Conversation
Including add a test case for this.
Also: make skinny_specs optional
{"sim_params.noise_abs": 8}, | ||
GridLocator( | ||
..., | ||
(..., ...), |
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.
What does ...
signify here in the GridLocator arguments? Is it meant to be taking all the metrics or is it empty?
Edit: I've read about the Ellipsis
object.
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.
Nice! Yeah, I wanted to borrow some notation from numpy that means "all elements". They're both what are called "singletons". Only one NoneType
object exists in a python program, just like only one EllipsisType
object ever exists.
I suppose I could have used None
instead. That may be an arguably better way, I just wanted to use an object that people associate with "all" of something.
), | ||
), | ||
"test-absrel3": _PlotPrefs( | ||
"absrel-newloc": _PlotPrefs( |
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.
For running the experiments, which one of the plot_prefs
do I use? test_absrel12
or absrel-newloc
?
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.
absrel-newloc
, but double check me that it's meaning is clear, given the GridLocator
argument
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.
LGTM
Substantial refactoring going on here. Big changes are (a) moving gridsearch to its own package to limit the coupling between it and the odes/pdes experiments, and (b) introducing the
gen_experiments.gridsearch.GridLocator
object, which allows one to specify criteria to match intermediate points in the gridsearch.To test it out, run any existing experiment, build a
GridLocator
object, and find the matching points withgridsearch.find_gridpoints(your_locator, results["plot_data"], results)