MVA Models manager
Introducing MVA Models Manager, inspired by Monitoring System ( !1483 (merged)), to simplify MVA models loading:
- MVA models are initialized inside algorithms, simplifying the model loading codes
- Only required models in a sequence would read data from .json files
- Provide an extendable framework to integrate gitlib2 in the future.
Was presented at 06 Aug WP2 meeting. Also adding documentation and examples of how new model can be created to Allen docs.
Example of usage:
#include "MultiLayerFCNN.cuh"
using NeuralNetworkType = Allen::MVAModels::MultiLayerNN;
NeuralNetworkType two_track_mva_nn {"two_track_mva_nn", "/allen_two_track_mva_model_June22.json"};
All existing MVA models are migrated to the framework in MR: SingleLayerFCNN (Matching Ghost Killer, Downstream Selector etc), MultiLayerFCNN (Two Track MVA, ElectronID NN), CatboostDesTree (MuonCatboostEvaluator).
Edited by Valerii Kholoimov
Merge request reports
Activity
added RTA label
added 3 commits
added 1 commit
- c03d923b - ParseAlgorithm changes to be able be runned via Moore
added 2 commits
added 12 commits
-
cf9d1b4c...9a669f1a - 4 commits from branch
2024-patches
- 67afd7f5 - introducing basic classes
- a7ab79f1 - moved all SLFCNN to updated version
- 2cbe3483 - comments cleaning
- 2068bb6c - adding TwoTrackMVA and ElectronID
- c63a87cc - MuonID plus CatboostMuonDT
- b38174bb - old code removing
- 7a7f4aa0 - ParseAlgorithm changes to be able be runned via Moore
- 76e2e643 - renaming
Toggle commit list-
cf9d1b4c...9a669f1a - 4 commits from branch
assigned to @vkholoim
- Resolved by Valerii Kholoimov
Please register or sign in to reply