This folder contains the ONNX models for the ParticleNetAK4 tagger designed for AK4 jet tagging (i.e., b-/c-tagging, q/g-tagging, PU jet ID). The model is trained on AK4CHS jets using RunIISummer19UL18MiniAOD the following samples:
/TTToSemiLeptonic_mtop171p5_TuneCP5_13TeV-powheg-pythia8/RunIISummer19UL18MiniAOD-106X_upgrade2018_realistic_v11_L1v1-v2/MINIAODSIM(300 files)/TTToSemiLeptonic_mtop173p5_TuneCP5_13TeV-powheg-pythia8/RunIISummer19UL18MiniAOD-106X_upgrade2018_realistic_v11_L1v1-v2/MINIAODSIM(300 files)/QCD_Pt_XtoY_TuneCP5_13TeV_pythia8/RunIISummer19UL18MiniAOD-PUForMUOVal_106X_upgrade2018_realistic_v11_L1v1-v3/MINIAODSIM(min_X=30, maxY=1400, 50 files each)/BulkGravitonToHHTo4Q_MX-600to6000_MH-15to250_part*_TuneCP5_13TeV-madgraph_pythia8/RunIISummer19UL18MiniAOD-multigridpack_106X_upgrade2018_realistic_v11_L1v1-v1/MINIAODSIM(300 files each)
The output classes of the DNN can be found in the preprocess.json file.