Machine learning workflows typically involve plenty of numerical computations in the form of mathematical and algebraic operations upon data stored as large vectors, matrices, or even tensors — matrix counterparts with three or more dimensions.



Source link