Colibri

Colibri is a reportengine application designed to perform Parton Distribution Function (PDF) fits using arbitrary parametrisations. It provides a flexible and adaptable open-source platform for PDF fitting.

Key Features:

  • Perform PDF fits using flexible parametrisations
  • Support for JAX-based computations (CPU by default)
  • Optional GPU (CUDA) support for accelerated JAX operations
  • Command-line scripts for common workflows (colibri, evolve_fit, etc.)
  • Integration with external PDF model repositories

Colibri is part of the PBSP project and is built to work with the NNPDF framework.

Installation:

# Via conda
conda env create -f environment.yml

# Or via pip
pip install git+https://github.com/HEP-PBSP/colibri.git

Citation guidelines: Please cite NNPDF (Eur. Phys. J. C 81 (2021) 958) and reportengine when using this code.

Languages: Python (97.5%), TeX (1.9%)

License: GPL-3.0

Elie Hammou
Elie Hammou
Postdoctoral researcher in Particle Physics

My research interests include (Beyond) the Standard Model physics, Collider Physics and Effective Field Theories.