DNN-based Tau-Id discrimians #25016
Merged
DNN-based Tau-Id discrimians #25016
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
- Defined base class for deep tau discriminators. - Removed weight files from home cms repository. Now using weights from cms-data. - Defined WP for both discriminators. Now all discriminators return the corresponding WP results. - Removed cfi files. Using fillDescriptions instead. - General code review and cleaning.
…ection with the new Tau-Ids
DNN tau IDs for CMSSW_10_4_X
|
The code-checks are being triggered in jenkins. |
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-25016/6991 |
MRD2F
added a commit
to MRD2F/cmssw
that referenced
this pull request
Dec 3, 2018
…de in the commit 194a1d5 from the PR cms-sw#25016 - RecoTauTag/RecoTau/plugins/DeepTauId.cc: code cleaning
|
please test with cms-sw/cmsdist#4554 |
cmsbuild
added
comparison-pending
tests-pending
and removed
comparison-available
tests-approved
labels
|
The tests are being triggered in jenkins. |
|
+1 The following merge commits were also included on top of IB + this PR after doing git cms-merge-topic: |
|
Comparison job queued. |
|
Comparison is ready The workflows 1001.0, 1000.0, 140.53, 136.85, 136.8311, 136.7611, 136.731, 4.22 have different files in step1_dasquery.log than the ones found in the baseline. You may want to check and retrigger the tests if necessary. You can check it in the "files" directory in the results of the comparisons Comparison Summary:
|
|
+1 |
forthommel
pushed a commit
to forthommel/cmssw
that referenced
this pull request
Dec 10, 2018
…_DNNTauIDs DNN-based Tau-Id discrimians
fabiocos
pushed a commit
that referenced
this pull request
Dec 12, 2018
* First implementation of deep tau id. * Building dpf isolation module * Adding in v1 * Adding in runTauIDMVA for other users * making things fully reproducible * Reorganisation of configuration files: cff split to cfi and cff * Some code cleaning * adapt to cfi/cff reorganization * Review of DPF and DeepTauId code. - Defined base class for deep tau discriminators. - Removed weight files from home cms repository. Now using weights from cms-data. - Defined WP for both discriminators. Now all discriminators return the corresponding WP results. - Removed cfi files. Using fillDescriptions instead. - General code review and cleaning. * Added example of a python configuration file to produce pat::Tau collection with the new Tau-Ids * requested changes on runDeepTauIDsOnMiniAOD.py * Clean runTauIdMVA.py tool and test config to run tauIDs * Made DeepTauId and DPFIsolation thread-safe * Finish implement thread-safe requirements on DPFIsolation * Disable DPFTau_2016_v1 and issue some warnings * Remove assigning value of variable to itself * - Implemented on runTauIdMVA the option to work with new training files quantized - Added a new parameter 'version' on runTauIdMVA, used on DPFIsolation - Changes on DeepTauId to reduce memory consumption * - Implementation of global cache to avoid reloading graph for each thread and reduce the memory consuption - Creation of class DeepTauCache in DeepTauBase, in which now is created graph and session - Implementation of two new static methods inside the class DeepTauBase: initializeGlobalCache and globalEndJob. The graph and DeepTauCache object are created now inside initializeGlobalCache * Applied changes on DeepTauBase to allow load new training files using memory mapping * Implemented TauWPThreshold class. TauWPThreshold class parses WP cut string (or value) provided in the python configuration. It is needed because the use of the standard StringObjectFunction class to parse complex expression results in an extensive memory usage (> 100 MB per expression). * Remove the qm.pb input files and leaving just the quantized and the original files * -Overall, changes to improve memory usage, among these are: - Implementation of global cache to avoid reloading graph for each thread - Creation of two new static methods inside the class DeepTauBase: initializeGlobalCache and globalEndJob. The graph and DeepTauCache object are created now inside initializeGlobalCache. The memory consumption of initializeGlobalCache for the original, quantized and files that are load using memory mapping method are in the memory_usage.pdf file - Implemented configuration to use new training files quantized, and set them as default - Implementation of configuration for load files using memory mapping. In our case there wasn't any improvement, respect at the memory consumption of this method, respect the quantized files, so this is not used, but set for future training files - General code review and cleaning. * Applied style comments * Applied style comments * Applied comments * Change to be by default the original training file for deepTau, instead of the quantized * Changes regarding forward-porting DNN-related developments from the PRs #105 and #106 from 94X to 104X * Applied commets of previus PR * cleaning code * Modification in the config to work with new label in files * Applied comment about the expected format of name of training file * Fix in last commit * Applied last comments * Changes regarding forward-porting DNN-related developments from the PRs #105 and #106 from 94X to 104X * Applied @perrotta comments on 104X * Fix error * Applied comments * Applied comments * Fix merge problem * Applied a few commets * Applied more changes * Applied a few small followups * Fixed error on DPFIsolation * Update DPFIsolation.cc * - RecoTauTag/RecoTau/plugins/DeepTauId.cc: Remove ' clusterVariables 'as a class member - RecoTauTag/RecoTau/test/runDeepTauIDsOnMiniAOD.py: Update globaltag and sample * Added changes in RecoTauTag/RecoTau/python/tools/runTauIdMVA.py made in the commit 194a1d5 from the PR #25016 * Fix error on runDeepTauIDsOnMiniAOD * Change the GT in RecoTauTag/RecoTau/test/runDeepTauIDsOnMiniAOD.py
fabiocos
pushed a commit
to fabiocos/cmssw
that referenced
this pull request
Jan 10, 2019
Building dpf isolation module Adding in v1 Adding in runTauIDMVA for other users making things fully reproducible Reorganisation of configuration files: cff split to cfi and cff Some code cleaning adapt to cfi/cff reorganization Review of DPF and DeepTauId code. - Defined base class for deep tau discriminators. - Removed weight files from home cms repository. Now using weights from cms-data. - Defined WP for both discriminators. Now all discriminators return the corresponding WP results. - Removed cfi files. Using fillDescriptions instead. - General code review and cleaning. Added example of a python configuration file to produce pat::Tau collection with the new Tau-Ids requested changes on runDeepTauIDsOnMiniAOD.py Clean runTauIdMVA.py tool and test config to run tauIDs Made DeepTauId and DPFIsolation thread-safe Finish implement thread-safe requirements on DPFIsolation Disable DPFTau_2016_v1 and issue some warnings - Implemented on runTauIdMVA the option to work with new training files quantized - Added a new parameter 'version' on runTauIdMVA, used on DPFIsolation - Changes on DeepTauId to reduce memory consumption - Implementation of global cache to avoid reloading graph for each thread and reduce the memory consuption - Creation of class DeepTauCache in DeepTauBase, in which now is created graph and session - Implementation of two new static methods inside the class DeepTauBase: initializeGlobalCache and globalEndJob. The graph and DeepTauCache object are created now inside initializeGlobalCache Applied changes on DeepTauBase to allow load new training files using memory mapping Implemented TauWPThreshold class. TauWPThreshold class parses WP cut string (or value) provided in the python configuration. It is needed because the use of the standard StringObjectFunction class to parse complex expression results in an extensive memory usage (> 100 MB per expression). Remove the qm.pb input files and leaving just the quantized and the original files -Overall, changes to improve memory usage, among these are: - Implementation of global cache to avoid reloading graph for each thread - Creation of two new static methods inside the class DeepTauBase: initializeGlobalCache and globalEndJob. The graph and DeepTauCache object are created now inside initializeGlobalCache. The memory consumption of initializeGlobalCache for the original, quantized and files that are load using memory mapping method are in the memory_usage.pdf file - Implemented configuration to use new training files quantized, and set them as default - Implementation of configuration for load files using memory mapping. In our case there wasn't any improvement, respect at the memory consumption of this method, respect the quantized files, so this is not used, but set for future training files - General code review and cleaning. Applied style comments Applied style comments Applied comments Change to be by default the original training file for deepTau, instead of the quantized Applied commets of previus PR cleaning code Modification in the config to work with new label in files Applied comment about the expected format of name of training file Fix in last commit Applied last comments Remove assigning value of variable to itself Applied @perrotta comments on 104X Fix error Applied comments Applied comments Fix merge problem Fix cherry-pick issue Applied a few commets Applied more changes Applied a few small followups Fixed error on DPFIsolation Added clamp function, that only is called it only when the std::clamp is not defined Update DPFIsolation.cc - RecoTauTag/RecoTau/test/runDeepTauIDsOnMiniAOD.py: Added changes made in the commit 194a1d5 from the PR cms-sw#25016 - RecoTauTag/RecoTau/plugins/DeepTauId.cc: code cleaning
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.
This pull request provides two new DNN-based Tau-Ids, DeepTau and DPFTau to be produced for pat::Taus with MiniAOD.
Details:
edm::stream::EDProducers calledDeepTauIdandDPFIsolationwith a common base classDeepTauBase,The new Tau-Ids are not part of any current workflow and primary they are meant to be run within users' workflows. Their integration as a part of NanoAOD or/and MiniAOD will be a subject of separate pull requests when this one is accepted.
Backports to 94X and 102X release series (used for 2016/17 and 2018 datasets) is foreseen when this one is accepted.
Recipe for tests (as new modules are not part of standard workflows):