carl-torch
PyTorch toolbox for multivariate Monte Carlo reweighting
carl-torch is a PyTorch toolbox for multivariate reweighting of Monte Carlo simulation, building on the CARL (Calibrated Approximate Ratio of Likelihoods) approach.
It supports training neural likelihood-ratio estimators and applying them to reweight events between simulation configurations — a practical building block for ATLAS modelling studies and synthetic-data workflows.