Change DerelativizeTransform to not use model predictions when use_raw_status_quo is True or when the status quo is infeasible.
#2036
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D51690727 |
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #2036 +/- ##
=======================================
Coverage 94.52% 94.52%
=======================================
Files 460 460
Lines 44460 44466 +6
=======================================
+ Hits 42025 42031 +6
Misses 2435 2435 ☔ View full report in Codecov by Sentry. |
|
This pull request was exported from Phabricator. Differential Revision: D51690727 |
dme65
pushed a commit
to dme65/Ax
that referenced
this pull request
Dec 1, 2023
…aw_status_quo` is `True` or when the status quo is infeasible. (facebook#2036) Summary: Pull Request resolved: facebook#2036 The was previously ignoring `use_raw_status_quo` and was using the values predicted by the model unless this errored out. This is misleading since we shouldn't be using the model predictions when `use_raw_status_quo` is `True`. This also resulted in weird behavior in the case where the status quo arm was within the search space bounds but didn't satisfy the parameter constraints. This transform would then use the model to predict the metrics of the status quo (ignoring that `use_raw_status_quo` was `True`), but the model wasn't trained on the status quo since it didn't satisfy the constraints. Thus, this resulted in the model having to extrapolate and producing very weird predictions. This diff changes the behavior to only use the model to predict the status quo metrics when (1) `use_raw_status_quo` is `False` and (2) the status quo is actually feasible. Reviewed By: Balandat Differential Revision: D51690727 fbshipit-source-id: 472e0b9462f6d5efe5554dd740a64146ee198d8f
dme65
force-pushed
the
export-D51690727
branch
from
c4cead9 to
0b423df
Compare
…aw_status_quo` is `True` or when the status quo is infeasible. (facebook#2036) Summary: Pull Request resolved: facebook#2036 The was previously ignoring `use_raw_status_quo` and was using the values predicted by the model unless this errored out. This is misleading since we shouldn't be using the model predictions when `use_raw_status_quo` is `True`. This also resulted in weird behavior in the case where the status quo arm was within the search space bounds but didn't satisfy the parameter constraints. This transform would then use the model to predict the metrics of the status quo (ignoring that `use_raw_status_quo` was `True`), but the model wasn't trained on the status quo since it didn't satisfy the constraints. Thus, this resulted in the model having to extrapolate and producing very weird predictions. This diff changes the behavior to only use the model to predict the status quo metrics when (1) `use_raw_status_quo` is `False` and (2) the status quo is actually feasible. Reviewed By: Balandat Differential Revision: D51690727 fbshipit-source-id: 3f2648a0091c88e5e4c9c2f388421d3f563f6372
|
This pull request was exported from Phabricator. Differential Revision: D51690727 |
dme65
force-pushed
the
export-D51690727
branch
from
0b423df to
470811f
Compare
|
This pull request has been merged in ef3b1c1. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
The was previously ignoring
use_raw_status_quoand was using the values predicted by the model unless this errored out. This is misleading since we shouldn't be using the model predictions whenuse_raw_status_quoisTrue.This also resulted in weird behavior in the case where the status quo arm was within the search space bounds but didn't satisfy the parameter constraints. This transform would then use the model to predict the metrics of the status quo (ignoring that
use_raw_status_quowasTrue), but the model wasn't trained on the status quo since it didn't satisfy the constraints. Thus, this resulted in the model having to extrapolate and producing very weird predictions.This diff changes the behavior to only use the model to predict the status quo metrics when (1)
use_raw_status_quoisFalseand (2) the status quo is actually feasible.Reviewed By: Balandat
Differential Revision: D51690727