Diffusion-based Model for Resonand Anomaly Detection

To train the model you can run:

python train.py

To sample from the trained model you can run:

python plot_jet.py  --sample --SR

The --SR flags determines if the background sample is done in the signal region (with --SR flag) or in the side-band region (without the --SE flag).

Drop the --sample to reproduce the plots without having to regenerate new files.

A classifier used to separate generated samples from data can be trained using:

python classify.py --SR

Where the --SR flag again determines where the backgrounds are sampled from.