tfm.optimization.math.fsum

Return an accurate floating point sum of values in the iterable seq.

Assumes IEEE-754 floating point arithmetic.