LHCb PID Compression

In this repository, we compress PID data via autoencoder

Pipeline description

First of all, we split data on train/validation/test and train autoencoder and xgboost

Then, we train GAN on compressed data (with given PIDs) and generate new data

Finally, we compute all accuracies

How to run

1. Create ./data/ and ./data/input/ folder
2. Add data.csv to ./data/input/
3. Configure config.py
4. Run run_pipeline.sh from ./scripts/
5. Collect results from ./data/output/