Add update for Tau NN with pT calibration

[Duchstf]

PR description:

• This PR extends on the last PR of updated Tau training to include pT calibration.
• All the results including efficiencies/rate are reported in this talk: https://indico.cern.ch/event/1321399/#17-phase-2-nn-puppi-taus
• The most current DP Notes include the updated tau training which has been merged: Retrained L1 Tau Trigger Integration #41492
• Updated DP Notes with pT calibration results coming soon after this PR is merged and verified.

@aloeliger I was trying to make another PR to l1t-offline cmssw repo but there is no CMSSW_14x versions yet, so I'm making the PR directly here, in principle though they should be the same changes (it doesn't want to automatically merge if I make a PR to cms-l1t-offline:phase2-l1t-integration-13_3_0_pre3. I assume this is because

PR validation:

• In progress, we are planning to use the producers from here: https://github.com/p2l1pfp/FastPUPPI/tree/14_0_X/NtupleProducer
• Everything pushed here should compile though.

[cmsbuild]

@Duchstf, CMSSW_14_0_X branch is closed for direct updates. cms-bot is going to move this PR to master branch.
In future, please use cmssw master branch to submit your changes.

[cmsbuild]

cms-bot internal usage

[cmsbuild]

-code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-43639/38295

ERROR: Build errors found during clang-tidy run.

L1Trigger/Phase2L1ParticleFlow/interface/taus/nnet_utils/nnet_dense.h:38:6: note: candidate template ignored: substitution failure [with data_T = layer13_t, res_T = layer14_t, CONFIG_T = config14]: array size is not a constant expression
--
L1Trigger/Phase2L1ParticleFlow/src/taus/TauNNIdHW.cc:76:3: error: no matching function for call to 'relu' [clang-diagnostic-error]
   76 |   nnet::relu<layer14_t, layer16_t, relu_config16>(layer14_out, layer16_out); // relu_5
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
L1Trigger/Phase2L1ParticleFlow/interface/taus/nnet_utils/nnet_activation.h:30:62: note: candidate template ignored: substitution failure [with data_T = layer14_t, res_T = layer16_t, CONFIG_T = relu_config16]: array size is not a constant expression
--
L1Trigger/Phase2L1ParticleFlow/src/taus/TauNNIdHW.cc:79:3: error: no matching function for call to 'dense' [clang-diagnostic-error]
   79 |   nnet::dense<layer16_t, layer17_t, config17>(layer16_out, layer17_out, w17, b17); // Dense_6
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
L1Trigger/Phase2L1ParticleFlow/interface/taus/nnet_utils/nnet_dense.h:38:6: note: candidate template ignored: substitution failure [with data_T = layer16_t, res_T = layer17_t, CONFIG_T = config17]: array size is not a constant expression
--
L1Trigger/Phase2L1ParticleFlow/src/taus/TauNNIdHW.cc:82:3: error: no matching function for call to 'sigmoid' [clang-diagnostic-error]
   82 |   nnet::sigmoid<layer17_t, result_t, sigmoid_config19>(layer17_out, layer19_out); // jetID_output
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
L1Trigger/Phase2L1ParticleFlow/interface/taus/nnet_utils/nnet_activation.h:61:6: note: candidate template ignored: substitution failure [with data_T = layer17_t, res_T = result_t, CONFIG_T = sigmoid_config19]: array size is not a constant expression
--
L1Trigger/Phase2L1ParticleFlow/src/taus/TauNNIdHW.cc:85:3: error: no matching function for call to 'dense' [clang-diagnostic-error]
   85 |   nnet::dense<layer16_t, result_t, config20>(layer16_out, layer20_out, w20, b20); // pT_output
      |   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
L1Trigger/Phase2L1ParticleFlow/interface/taus/nnet_utils/nnet_dense.h:38:6: note: candidate template ignored: substitution failure [with data_T = layer16_t, res_T = result_t, CONFIG_T = config20]: array size is not a constant expression
--
gmake: *** [config/SCRAM/GMake/Makefile.coderules:129: code-checks] Error 2
gmake: *** [There are compilation/build errors. Please see the detail log above.] Error 2

[Duchstf]

Hmm I'm not sure what's happening here, the PR compiles locally for me and I checked all the files in.

[cmsbuild]

+code-checks

Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-43639/38296

• This PR adds an extra 48KB to repository

• Found files with invalid states:

  • L1Trigger/Phase2L1ParticleFlow/interface/taus/nnet_utils/nnet_dense_latency.h:
    • Added: d48b499
    • Deleted: fb60584
  • L1Trigger/Phase2L1ParticleFlow/interface/taus/nnet_utils/nnet_types.h:
    • Added: d48b499
    • Deleted: fb60584
  • L1Trigger/Phase2L1ParticleFlow/interface/taus/nnet_utils/nnet_helpers.h:
    • Added: d48b499
    • Deleted: fb60584

[cmsbuild]

A new Pull Request was created by @Duchstf (Duc Hoang) for master.

It involves the following packages:

• L1Trigger/Phase2L1ParticleFlow (upgrade, l1)

@epalencia, @subirsarkar, @cmsbuild, @srimanob, @aloeliger can you please review it and eventually sign? Thanks.
@Martin-Grunewald, @missirol this is something you requested to watch as well.
@rappoccio, @antoniovilela, @sextonkennedy you are the release manager for this.

cms-bot commands are listed here

[aloeliger]

Hi @Duchstf I am starting to catch up on things after the holidays.

@aloeliger I was trying to make another PR to l1t-offline cmssw repo but there is no CMSSW_14x versions yet, so I'm making the PR directly here, in principle though they should be the same changes (it doesn't want to automatically merge if I make a PR to cms-l1t-offline:phase2-l1t-integration-13_3_0_pre3. I assume this is because

If you need we can bump the IB version forward. We're trying to be a bit more flexible about that and if any developers want to, we can integrate the latest changes here in CMSSW.

On another note, are there any plans to change this emulator into a more standard hls4ml emulator style? At the recent ML@L1 workshop Sioni made a tutorial on this, and a large part of my talk is on the process with some examples.

[Duchstf]

@aloeliger Hello, thank you for the comment!! And sorry for the late reply (I just came back from holiday as well ...) Yes we are definitely planning to follow the tutorial and implement it that way!! I was a bit busy with qualifying exams right before the deadline so I couldn't look into it. And to meet the PR deadline I basically did what I know how to do best.

But yea we might push more commits later that follow the tutorials if you prefer that.

[epalencia]

Please test

[cmsbuild]

+1

Summary: https://cmssdt.cern.ch/SDT/jenkins-artifacts/pull-request-integration/PR-8881a0/36793/summary.html
COMMIT: fb60584
CMSSW: CMSSW_14_0_X_2024-01-10-1100/el8_amd64_gcc12
User test area: For local testing, you can use /cvmfs/cms-ci.cern.ch/week1/cms-sw/cmssw/43639/36793/install.sh to create a dev area with all the needed externals and cmssw changes.

Comparison Summary

Summary:

• You potentially added 3 lines to the logs
• Reco comparison results: 19 differences found in the comparisons
• DQMHistoTests: Total files compared: 48
• DQMHistoTests: Total histograms compared: 3247277
• DQMHistoTests: Total failures: 1216
• DQMHistoTests: Total nulls: 0
• DQMHistoTests: Total successes: 3246039
• DQMHistoTests: Total skipped: 22
• DQMHistoTests: Total Missing objects: 0
• DQMHistoSizes: Histogram memory added: 0.0 KiB( 47 files compared)
• Checked 200 log files, 161 edm output root files, 48 DQM output files
• TriggerResults: no differences found

[aloeliger]

@epalencia I think it's fine to sign on this for now and the overhaul towards the other hlsrml emulators can wait. Anything from your side?

[epalencia]

+l1

[epalencia]

ping @cms-sw/upgrade-l2

[srimanob]

+Upgrade

[cmsbuild]

This pull request is fully signed and it will be integrated in one of the next master IBs (tests are also fine). This pull request will now be reviewed by the release team before it's merged. @sextonkennedy, @rappoccio, @antoniovilela (and backports should be raised in the release meeting by the corresponding L2)

[rappoccio]

+1