Interesting to note that decimal fractions (e.g. 0.3) are awkward to represent with floats in computers, whereas fractions based on 1/2, 1/4, 1/8, etc. that are often used in the non-metric system are a very natural fit for the float representation.
If you use decimal with metric, the units and prefixes play along with you very well (huge advantage). If you use other fractions with metric, you get no special bonus or burden (neutral).
If you use decimal with US customary ("imperial"), you get no special bonus or burden (netural). If you use other fractions with US customary, you get no special bonus or burden (neutral).
In summary, there is absolutely no rational reason to stick with US customary, besides legacy compatibility (technical debt).