Thanks to visit codestin.com
Credit goes to Github.com

Skip to content

Oguz-Guzel/conditional-flow-matching

 
 

Repository files navigation

Using TorchCFM: a Conditional Flow Matching library for background estimation

WORK IN PROGRESS

Check out the original repo here for more functionality here : https://github.com/atong01/conditional-flow-matching

We use the tabular data generation jupyter notebook located in examples/tabular/ to demonstrate how to use TorchCFM for generation of background events, specifically the Drell-Yan process.

The generated distributions include $p_T$, $\eta$ and $\phi$ of the leptons, as well as the invariant mass of the dilepton system.

examples/tabular/feature_distributions.png

examples/tabular/mll_category_distribution.png

The invariant mass calculated from the generated distribution:

examples/tabular/mll_from_leptons_distribution.png

About

TorchCFM: a Conditional Flow Matching library

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 99.8%
  • Shell 0.2%