Phase2 L1 NN Tau #30505
Closed
Phase2 L1 NN Tau #30505
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
|
assign upgrade |
|
The code-checks are being triggered in jenkins. |
|
assign l1 |
|
-code-checks ERROR: Build errors found during clang-tidy run. |
kpedro88
reviewed
Jul 2, 2020
| void addTau(l1t::PFCandidate &iCand, | ||
| const l1t::PFCandidateCollection &iParts, | ||
| std::unique_ptr<PFTauCollection> &outputTaus); | ||
| float deltaR(auto iPart1, auto iPart2); |
There was a problem hiding this comment.
There was a problem hiding this comment.
(also, abbreviated function templates using auto are a C++20 feature, not supported in CMSSW yet, hence the compilation error in the PR test)
| #ifndef L1TRIGGER_PHASE2L1PARTICLEFLOW_L1NNTAU_H | ||
| #define L1TRIGGER_PHASE2L1PARTICLEFLOW_L1NNTAU_H | ||
|
|
||
| #include <iostream> |
|
|
||
| #include <iostream> | ||
| #include <vector> | ||
| #include <TLorentzVector.h> |
| @@ -0,0 +1,30 @@ | |||
| #ifndef L1TRIGGER_PHASE2L1PARTICLEFLOWS_TAUNNID_H | |||
| #define L1TRIGGER_PHASE2L1PARTICLEFLOWS_TAuNNID_H | |||
There was a problem hiding this comment.
inconsistent flags, please check spelling/capitalization
| fNParticles_(cfg.getParameter<int>("nparticles")), | ||
| fL1PFToken_(consumes<vector<l1t::PFCandidate> >(cfg.getParameter<edm::InputTag>("L1PFObjects"))) { | ||
| std::string lNNFile = cfg.getParameter<std::string>("NNFileName"); //,"L1Trigger/Phase2L1Taus/data/tau_3layer.pb"); | ||
| fTauNNId = new TauNNId(); |
There was a problem hiding this comment.
memory leak, please use smart pointer
also, member variable should end in underscore
| #include "L1Trigger/Phase2L1ParticleFlow/interface/TauNNId.h" | ||
| #include <iostream> | ||
| #include <cmath> | ||
| #include <TMath.h> |
| delete graphDef_; | ||
| } | ||
| void TauNNId::initialize(const std::string &iInput, const std::string &iWeightFile, int iNParticles) { | ||
| std::string cmssw_base_src = getenv("CMSSW_BASE"); |
There was a problem hiding this comment.
better to use FileInPath to get the full path automatically
| fEta_ = std::make_unique<float>(fNParticles_); | ||
| fPhi_ = std::make_unique<float>(fNParticles_); | ||
| fId_ = std::make_unique<float>(fNParticles_); | ||
| fInput_ = iInput; //tensorflow::run(session_, { { "input_1:0",input } }, { "dense_4/Sigmoid:0" }, &outputs); |
| NNvectorVar_.push_back(0); | ||
| continue; | ||
| } | ||
| fId_.get()[i0] == l1t::PFCandidate::Photon ? NNvectorVar_.push_back(1) : NNvectorVar_.push_back(0); //Photon |
There was a problem hiding this comment.
simplify: NNvectorVar_.push_back(fId_.get()[i0] == l1t::PFCandidate::Photon);
(also for lines below)
| fPt_.get()[i0] = iParts[i0].pt(); | ||
| fEta_.get()[i0] = iSeed.eta() - iParts[i0].eta(); | ||
| float lDPhi = iSeed.phi() - iParts[i0].phi(); | ||
| if (lDPhi > TMath::Pi()) |
There was a problem hiding this comment.
|
The code-checks are being triggered in jenkins. |
|
-code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-30505/16702
Code check has found code style and quality issues which could be resolved by applying following patch(s)
|
|
addressed cmments and now rebasing off of #30510 |
rekovic
force-pushed
the
pr111x-L1ParticleFlow-Phase2-clean-with-Tau
branch
from
235c856
to
7eed165
Compare
|
The code-checks are being triggered in jenkins. |
|
+code-checks Logs: https://cmssdt.cern.ch/SDT/code-checks/cms-sw-PR-30505/16727
|
|
Comparison is ready Comparison Summary:
|
|
+1 |
makortel
reviewed
Jul 7, 2020
|
|
||
| private: | ||
| tensorflow::Session *session_; | ||
| tensorflow::GraphDef *graphDef_; |
There was a problem hiding this comment.
How big is the graph? I think we've typically shared the graph across the stream modules in an edm::GlobalCache.
There was a problem hiding this comment.
also, a smart pointer could be used (though it seems like that isn't common practice for CMSSW TF code at the moment, something to be improved globally)
kpedro88
reviewed
Jul 7, 2020
| return; | ||
| } | ||
| pfChargedHadrons_seeds.push_back(pfChargedHadrons_sort[0]); | ||
| for (unsigned int i0 = 1; i0 < pfChargedHadrons_sort.size(); i0++) { |
| pfChargedHadrons_seeds.push_back(pfChargedHadrons_sort[0]); | ||
| for (unsigned int i0 = 1; i0 < pfChargedHadrons_sort.size(); i0++) { | ||
| bool pMatch = false; | ||
| for (unsigned int i1 = 0; i1 < pfChargedHadrons_seeds.size(); i1++) { |
| if (int(pfChargedHadrons_seeds.size()) > fMaxTaus_ - 1) | ||
| break; | ||
| } | ||
| for (unsigned int i0 = 0; i0 < pfChargedHadrons_seeds.size(); i0++) { |
| TLorentzVector lCand; | ||
| lCand.SetPtEtaPhiM(0, 0, 0, 0); | ||
| int lId = 0; | ||
| for (auto l1PFCand : iParts) { |
| l1t::PFTau l1PFTau(tempP4, NN, 0, lId); | ||
| outputTaus->push_back(l1PFTau); | ||
| } | ||
| float L1NNTauProducer::deltaR(const l1t::PFCandidate& iPart1, const l1t::PFCandidate& iPart2) { |
There was a problem hiding this comment.
| fMaxTaus_(cfg.getParameter<int>("maxtaus")), | ||
| fNParticles_(cfg.getParameter<int>("nparticles")), | ||
| fL1PFToken_(consumes<vector<l1t::PFCandidate> >(cfg.getParameter<edm::InputTag>("L1PFObjects"))) { | ||
| std::string lNNFile = cfg.getParameter<std::string>("NNFileName"); //,"L1Trigger/Phase2L1Taus/data/tau_3layer.pb"); |
|
|
||
| private: | ||
| tensorflow::Session *session_; | ||
| tensorflow::GraphDef *graphDef_; |
There was a problem hiding this comment.
also, a smart pointer could be used (though it seems like that isn't common practice for CMSSW TF code at the moment, something to be improved globally)
| fPt_.get()[i0] = iParts[i0].pt(); | ||
| fEta_.get()[i0] = iSeed.eta() - iParts[i0].eta(); | ||
| float lDPhi = iSeed.phi() - iParts[i0].phi(); | ||
| if (lDPhi > M_PI) |
There was a problem hiding this comment.
| } | ||
| std::sort(iParts.begin(), iParts.end(), [](l1t::PFCandidate i, l1t::PFCandidate j) { return (i.pt() > j.pt()); }); | ||
| for (unsigned int i0 = 0; i0 < iParts.size(); i0++) { | ||
| if (i0 > 10) |
There was a problem hiding this comment.
why is this hardcoded? is it the same as the "nparticles" python parameter? if so, just use that...
(also, just make the loop condition i0 < std::min(iParts.size(),fNparticles_))
Merged
|
@rekovic there is a conflict |
|
@rekovic there are also still comments to be addressed |
|
@rekovic what is the status of this PR |
This was referenced Jul 15, 2020
|
@rekovic ping |
This was referenced Jul 28, 2020
Closed
|
-1 |
|
@rekovic please let me know when you want to open this PR back |
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.
PR description:
This is effectively the PR #30482 (L1Trigger/Phase2L1ParticleFlow package which has been reviewed fully) augmented by the very last one commit dd150ad which adds the producer for L1 NN Taus, contained in the following five files:
The NNTau producer is added to the Phase2 L1T sequence in
@kpedro88
Please review those.
PR validation:
if this PR is a backport please specify the original PR and why you need to backport that PR:
Before submitting your pull requests, make sure you followed this checklist: