particleflow
particleflow copied to clipboard
jpata
Machine-learned, GPU-accelerated particle flow reconstruction
- updated the response plot definition in cmssw.ipynb - some fixes and updates to the transformer model - switch to using focal loss with logits (pre-softmax) (pointed out by Dylan)...
One of the fundamental questions is the computation of the loss function between two sets of particles of different multiplicity. For example, given the true set, with **no** natural ordering:...
As suggested by JK here: https://indico.cern.ch/event/1189908/contributions/5008659/attachments/2494293/4283744/20220818_jk_mlpf.pdf it might be useful to try to improve the reconstruction of event-level quantities by computing and comparing local averages of energies for each true/reconstructed...
- [x] get the basic model running on GPU with ONNXRuntime: https://github.com/jpata/cmssw/commit/36be715fa00457c310acae3c033f4788bd47a26b - [ ] run the model efficiently, with async executing and batching through e.g. SONIC: - update this...
For the moment, we concatenate input elements of different types (tracks, clusters, ...) into a single feature matrix. This means that there are features that may be defined for one...
In the `pipeline.py evaluate` step, we could produce some printouts or plots of badly reconstructed cases for further debugging.
Using this Pythia8 card: ``` Random:setSeed = on Main:numberOfEvents = 100 ! number of events to generate Main:timesAllowErrors = 5 ! how many aborts before run stops ! 2) Settings...
1. L1 attempts to correct for PU: simulate the same event with & without PU 2. L2L3 attempts to correct gen to reco: match genjets to recojets, derive a correction...
- PyTorch-based training on Gaudi HPU Issue when importing `torch_cluster` ```bash >>> import torch_cluster Traceback (most recent call last): File "", line 1, in File "/usr/local/lib/python3.10/dist-packages/torch_cluster/__init__.py", line 18, in torch.ops.load_library(spec.origin)...
