|
BmnRoot 26.10.0-18-g4e3414f9
Software framework of the BM@N experiment
|
Solution of the alignment problem of the new alignment by mean squared error minimization with neural network training techniques: for each alignable detector element a custom single-layer model containing the measurement model is trained, and its weights give the alignment parameters.
| File | Description |
|---|---|
| expdataNN.C | Exports the hits of the experimental data for the training |
| train.py | Training of the models with PyTorch (dependencies are declared in the script header for uv) |
| draw_tracks.py | Drawing of the tracks with matplotlib |
| out.txt | Example output of the training |
Files | |
| expdataNN.C | |