Fabber is a Bayesian model-fitting framework designed to fit nonlinear parameterised models to timeseries data, particularly 4D fMRI data such as ASL, CEST, DCE, DSC, etc.
PYFAB offers a Python interface to the tool which can work with the command line Fabber tool, or the shared library version, depending on which is installed.
Fabber must be installed before use - See https://fabber_core.readthedocs.io/. Fabber and a selection of models will be available in the upcoming FSL 6.0.1 release
Full documentation at https://pyfab.readthedocs.io/