Add HLT_DoubleMediumPFPuppiParTTauh30_eta2p1 path - #51705
Open
agrubercms wants to merge 2 commits into
Open
Conversation
Contributor
|
cms-bot internal usage |
Contributor
|
-code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51705/50596 Code check has found code style and quality issues which could be resolved by applying following patch(s)
|
agrubercms
force-pushed
the
GloParT_clean_PR
branch
from
August 14, 2026 10:29
11b9638 to
2858f55
Compare
Contributor
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51705/50597 |
Contributor
|
Pull request #51705 was updated. |
agrubercms
force-pushed
the
GloParT_clean_PR
branch
from
August 14, 2026 11:55
2858f55 to
73a79fc
Compare
Contributor
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51705/50599 |
Contributor
|
Pull request #51705 was updated. |
mmusich
reviewed
Aug 14, 2026
mmusich
left a comment
mmusich
left a comment
Contributor
There was a problem hiding this comment.
can the new path be added also to the timing menu?
Also can the UParT sequence be added to the NGT scouting path?
Contributor
|
test parameters:
|
cmsbuild
added
dqm-pending
tests-pending
code-checks-pending
and removed
tests-rejected
code-checks-approved
labels
Contributor
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-51705/50723
|
Contributor
|
@cmsbuild, please test |
cmsbuild
added
tests-started
code-checks-approved
and removed
tests-pending
code-checks-pending
labels
Contributor
|
Pull request #51705 was updated. |
Contributor
|
+1 Size: This PR adds an extra 52KB to repository HLT P2 Timing: chart Comparison SummarySummary:
AMD_MI300X Comparison SummarySummary:
AMD_W7900 Comparison SummarySummary:
NVIDIA_H100 Comparison SummarySummary:
NVIDIA_L4 Comparison SummarySummary:
NVIDIA_L40S Comparison SummarySummary:
NVIDIA_T4 Comparison SummarySummary:
Max Memory Comparisons exceeding threshold NVIDIA_L4@cms-sw/core-l2 , I found 1 workflow step(s) with memory usage exceeding the error threshold: Expand to see workflows ...
|
Contributor
|
@agrubercms I am ready to sign for HLT. Is there any reason to keep this "in progress"? |
Contributor
Author
Hi @mmusich, we're planning to present this work next week at HLT Upgrade, and I thought I'd wait for that in case there's substantial comments - but if you're okay with it already, I'm happy to remove the "draft" label |
Contributor
if it gets undrafted now, you'll get a chance to get reviews from the other involved groups, so I warmly recommend you to do that. |
agrubercms
marked this pull request as ready for review
August 26, 2026 13:05
Contributor
|
@cmsbuild, ping |
Contributor
|
A new Pull Request was created by @agrubercms for master. It involves the following packages:
@Martin-Grunewald, @Moanwar, @ctarricone, @gabrielmscampos, @jfernan2, @mandrenguyen, @mmusich, @rseidita, @srimanob can you please review it and eventually sign? Thanks. cms-bot commands are listed here |
Contributor
|
Jut to document for the record, the results from the timing benchmark executed by the bot (and reported at #51705 (comment)):
the overall cost is 5512.5 ms/ev - 5447.6 ms/ev = 64 ms/ev, (i.e. an increase of ~1%). The diff is the following. I think this is spoiled mostly by migration from and to I/O 
the actual additional modules are pretty small:
Resource increase usage is negligible charts |
Contributor
|
+hlt for #51705 at agrubercms@cc010d4;
|
Member
|
+dqm |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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.
PR description:
This PR introduces a new Phase-2 tau HLT path (HLT_DoubleMediumPFPuppiParTTauh30_eta2p1) based on a ParticleTransformer model. The model (+documentation) can be found in this PR to the cms-data repository. It was developed to also be able to be used for b-tagging, but was not yet tested for this purpose - to illustrate this, the PR includes b- and c- discriminators, which are currently commented out. Most implementation choices were made to maintain compatibility with the existing tools used by the BTV group (most notably the hltupler framework to produce the training ntuples).
Development progress and measurements were presented regularly in Tau internal meetings, e.g. here and here.
The new model significantly outperforms the existing DeepTau model across all evaluated metrics - these are tau tagging efficiency, trigger rate and computational runtime.
PR validation:
The standard tests have been run:
scram b runtests use-ibeosrunTheMatrix.py -l limited -i all --ibeosModel performance has been evaluated using the same sample composition as was used in the training, but with a different subgroup of the samples to ensure evaluation is done on events not used in the model training.
To test this PR, the PR to cms-data (mentioned above) containing the onnx model would need to be included.