Update SOBOL transition criterion to excldue ABANDONED and FAILED trials in initalization budget #4776
+60
−3
Conversation
This file contains hidden or 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
meta-cla
bot
added
the
CLA Signed
|
@sunnyshen321 has exported this pull request. If you are a Meta employee, you can view the originating Diff in D90807058. |
sunnyshen321
pushed a commit
to sunnyshen321/Ax
that referenced
this pull request
Jan 15, 2026
…als in initalization budget (facebook#4776) Summary: Exclude ABANDONED and FAILED trials from initalization budget transition criterion f1020104666 has 3 abandoned trials (missing qps metrics) and 2 completed trials. But it hit a dead end: - Can't generate more Sobol trials (blocked by criterion 1 - 'block_gen_if_met': True) - Can't transition to BO (blocked by criterion 2 - not enough COMPLETED trials) `block_gen_if_met = True` is a sensible default as we don't necessarily want to spend more SOBOL trials just to meet the max_parallelism, but we should excldue FAILED/ABANDONED trials to prevent issues like f1020104666 This means that if we have 3 COMPLETED trials and 2 FAILED/ABANDONED trials, we will stay in SOBOL (previous behavior: move to BO). Intuitively it shouldn't make a big difference in optimization result as BO model is probably not good with 3 datapoints, and probably behaves like SOBOL anyway (except for BO is more likely to generate boundary points). Reviewed By: saitcakmak Differential Revision: D90807058
sunnyshen321
force-pushed
the
export-D90807058
branch
from
451977a to
70d7d48
Compare
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #4776 +/- ##
=======================================
Coverage 96.72% 96.72%
=======================================
Files 583 583
Lines 60880 60903 +23
=======================================
+ Hits 58887 58910 +23
Misses 1993 1993 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
…als in initalization budget (facebook#4776) Summary: Exclude ABANDONED and FAILED trials from initalization budget transition criterion f1020104666 has 3 abandoned trials (missing qps metrics) and 2 completed trials. But it hit a dead end: - Can't generate more Sobol trials (blocked by criterion 1 - 'block_gen_if_met': True) - Can't transition to BO (blocked by criterion 2 - not enough COMPLETED trials) `block_gen_if_met = True` is a sensible default as we don't necessarily want to spend more SOBOL trials just to meet the max_parallelism, but we should excldue FAILED/ABANDONED trials to prevent issues like f1020104666 This means that if we have 3 COMPLETED trials and 2 FAILED/ABANDONED trials, we will stay in SOBOL (previous behavior: move to BO). Intuitively it shouldn't make a big difference in optimization result as BO model is probably not good with 3 datapoints, and probably behaves like SOBOL anyway (except for BO is more likely to generate boundary points). Reviewed By: saitcakmak Differential Revision: D90807058
sunnyshen321
force-pushed
the
export-D90807058
branch
from
70d7d48 to
d64d7e0
Compare
|
This pull request has been merged in 3156389. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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:
Exclude ABANDONED and FAILED trials from initalization budget transition criterion
f1020104666 has 3 abandoned trials (missing qps metrics) and 2 completed trials. But it hit a dead end:
block_gen_if_met = Trueis a sensible default as we don't necessarily want to spend more SOBOL trials just to meet the max_parallelism, but we should excldue FAILED/ABANDONED trials to prevent issues like f1020104666This means that if we have 3 COMPLETED trials and 2 FAILED/ABANDONED trials, we will stay in SOBOL (previous behavior: move to BO). Intuitively it shouldn't make a big difference in optimization result as BO model is probably not good with 3 datapoints, and probably behaves like SOBOL anyway (except for BO is more likely to generate boundary points).
Differential Revision: D90807058