We won't know until after the votes are counted, and that's only assuming exit polls are correct. (questionable)
We can get clues from the turnout by demographic—where that information is available. The relevant information is not the total numbers, but the comparison to 2016.
Any analysis that fails to compare apples to apples, 2020 to 2016, is _useless_.
If the college educated independents in the cities and inner suburbs turn out in larger numbers than 2016, we are probably looking at Democratic votes.
If non college educated exurban and rural independents are turning out in larger numbers than 2016, we are probably looking at Republican votes.