Skip to content

Deterministic DRN #42950

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
cmsbuild merged 2 commits into from
Oct 9, 2023
Merged

Deterministic DRN #42950

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
30dc9c0
Producer with printouts
ssrothman Sep 13, 2023
6b48c07
code-format
ssrothman Oct 4, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions PhysicsTools/PatAlgos/python/slimming/patPhotonDRNCorrector_cfi.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,8 +7,8 @@
Client = patPhotonDRNCorrectionProducer.Client.clone(
mode = 'Async',
allowedTries = 1,
modelName = 'photonObjectEnsemble',
modelConfigPath = 'RecoEgamma/EgammaPhotonProducers/data/models/photonObjectEnsemble/config.pbtxt',
modelName = 'photonObjectCombined',
modelConfigPath = 'RecoEgamma/EgammaPhotonProducers/data/models/photonObjectCombined/config.pbtxt',
timeout = 10
)
)
24 changes: 12 additions & 12 deletions RecoEgamma/EgammaTools/plugins/DRNCorrectionProducerT.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -211,42 +211,42 @@ void DRNCorrectionProducerT<T>::acquire(edm::Event const& iEvent, edm::EventSetu
* gx: (rho, H/E) additional high-level features.
* batch{SB}: graph models require explicitely passing the particle index for each vertex
*/
auto& inputxECAL = iInput.at("xECAL");
auto& inputxECAL = iInput.at("xECAL__0");
inputxECAL.setShape(0, nHitsECAL);
auto dataxECAL = inputxECAL.allocate<float>();
auto& vdataxECAL = (*dataxECAL)[0];

auto& inputfECAL = iInput.at("fECAL");
auto& inputfECAL = iInput.at("fECAL__1");
inputfECAL.setShape(0, nHitsECAL);
auto datafECAL = inputfECAL.allocate<int64_t>();
auto& vdatafECAL = (*datafECAL)[0];

auto& inputGainECAL = iInput.at("gainECAL");
auto& inputGainECAL = iInput.at("gain__2");
inputGainECAL.setShape(0, nHitsECAL);
auto dataGainECAL = inputGainECAL.allocate<int64_t>();
auto& vdataGainECAL = (*dataGainECAL)[0];

auto& inputGx = iInput.at("gx");
auto& inputGx = iInput.at("graph_x__5");
inputGx.setShape(0, nValidPart_);
auto dataGx = inputGx.allocate<float>();
auto& vdataGx = (*dataGx)[0];

auto& inputBatchECAL = iInput.at("batchECAL");
auto& inputBatchECAL = iInput.at("xECAL_batch__6");
inputBatchECAL.setShape(0, nHitsECAL);
auto dataBatchECAL = inputBatchECAL.allocate<int64_t>();
auto& vdataBatchECAL = (*dataBatchECAL)[0];

auto& inputxES = iInput.at("xES");
auto& inputxES = iInput.at("xES__3");
inputxES.setShape(0, nHitsES);
auto dataxES = inputxES.allocate<float>();
auto& vdataxES = (*dataxES)[0];

auto& inputfES = iInput.at("fES");
auto& inputfES = iInput.at("fES__4");
inputfES.setShape(0, nHitsES);
auto datafES = inputfES.allocate<int64_t>();
auto& vdatafES = (*datafES)[0];

auto& inputBatchES = iInput.at("batchES");
auto& inputBatchES = iInput.at("xES_batch__7");
inputBatchES.setShape(0, nHitsES);
auto dataBatchES = inputBatchES.allocate<int64_t>();
auto& vdataBatchES = (*dataBatchES)[0];
Expand Down Expand Up @@ -346,6 +346,7 @@ void DRNCorrectionProducerT<T>::acquire(edm::Event const& iEvent, edm::EventSetu
/*
* Convert input tensors to server data format
*/

inputxECAL.toServer(dataxECAL);
inputfECAL.toServer(datafECAL);
inputGainECAL.toServer(dataGainECAL);
Expand All @@ -368,16 +369,15 @@ void DRNCorrectionProducerT<T>::produce(edm::Event& iEvent, const edm::EventSetu
//if there are no particles, the fromServer() call will fail
//but we can just put() an empty valueMap
if (nPart_) {
const auto& muOut = iOutput.at("mu").fromServer<float>();
const auto& sigmaOut = iOutput.at("sigma").fromServer<float>();
const auto& out = iOutput.at("combined_output__0").fromServer<float>();

unsigned i = 0;
float mu, sigma, Epred, sigmaPred, rawE;
for (unsigned iPart = 0; iPart < nPart_; ++iPart) {
const auto& part = particles_->at(iPart);
if (!skip(part)) {
mu = correction(muOut[0][0 + 6 * i]);
sigma = resolution(sigmaOut[0][0 + 5 * i]);
mu = correction(out[0][0 + 11 * i]);
sigma = resolution(out[0][6 + 11 * i]);
++i;

rawE = part.superCluster()->rawEnergy();
Expand Down