Update target trial selection logic to only consider trials with data for optimization config metrics #4742
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
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #4742 +/- ##
=======================================
Coverage 96.72% 96.72%
=======================================
Files 582 582
Lines 60748 60800 +52
=======================================
+ Hits 58758 58811 +53
+ Misses 1990 1989 -1 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 6, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
force-pushed
the
export-D90089411
branch
from
c7b8f93 to
7211102
Compare
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 7, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
force-pushed
the
export-D90089411
branch
from
7211102 to
54551ac
Compare
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 7, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
force-pushed
the
export-D90089411
branch
from
54551ac to
be89b7a
Compare
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 7, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
force-pushed
the
export-D90089411
branch
from
be89b7a to
65a540f
Compare
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 9, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
force-pushed
the
export-D90089411
branch
2 times, most recently
from
f39f5ba to
73f3439
Compare
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 13, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 13, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
force-pushed
the
export-D90089411
branch
from
73f3439 to
f0f0d2c
Compare
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 13, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
force-pushed
the
export-D90089411
branch
2 times, most recently
from
106c194 to
9e8508b
Compare
mgarrard
added a commit
to mgarrard/Ax
that referenced
this pull request
Jan 14, 2026
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
… for optimization config metrics (facebook#4742) Summary: This updates target trial selection logic in the following ways: 1. We only consider trials that either (a) have data for *all* opt config metrics or (b) have data for *all* metrics -- previously if a trial had data for some opt config metrics it passed the check, but this partial data setup causes issues downstream 2. If there is no long run trial, we fallback to short run instead of not identifying a target trial 3. filters out "stale" trials, ie trials that were completed over 10 days ago -- hitting this point would be pretty far down our priority list, but was an idea ItsMrLin initially had that i thought was really interesting 4. lastly if no trials exist after stale is filtered out, use the stale ones anyway (necessary for benchmarking) 5. if we still can't find anything, it will return none Reviewed By: ItsMrLin Differential Revision: D90089411 Privacy Context Container: L1307644
mgarrard
force-pushed
the
export-D90089411
branch
from
9e8508b to
217792a
Compare
|
This pull request has been merged in 2d69d98. |
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:
This updates target trial selection logic in the following ways:
Differential Revision:
D90089411
Privacy Context Container: L1307644