In general there is no problem, you even have a PEP that talks about how to do it in the right way: https://www.python.org/dev/peps/pep-0263/.
Of course, it does not free you from having problems because of some external tool that reads the code and is not able to handle this, to give problems when it opens in a bad text editor and other things. This has been addressed in general in Is there a problem in the use of Unicode characters for identifiers in code? I would not abuse, at the maximum I would use accents.
Skinny and elegant, they do not always go together (there are cases that do). I can not imagine how special characters can leave the code more elegant and at the same time, although elegant can be a bit subjective.