r-cran-numderiv binary package in Ubuntu Focal arm64

 This package provide methods for calculating (usually) accurate
 numerical first and second order derivatives. Accurate calculations
 are done using Richardson's extrapolation or, when applicable, a
 complex step derivative is available. A simple difference method is
 also provided. Simple difference is (usually) less accurate but is
 much quicker than Richardson's extrapolation and provides a useful
 cross-check. Methods are provided for real scalar and vector valued
 functions.

Source package