Bugfix/names for multiplications
This merge implements the following changes in the API:
-
inner()
,inner<N>()
->contractInner()
,contractInner<N>()
, contraction over just the inner-most or theN
inner-most indices. -
frobenius()
->frobenius2()
, square of the (generalized for tensors) frobenius norm, i.e. the square-sum over all components. - NOTHING ->
inner()
, contraction over all indices, i.e. the generalized scalar product for tensors. Their signature has to match.
Further, operator*()
performs contractInner<1>()
which yields the usual scalar-product for vectors and matrix-matrix and matrix-vector, vector-matrix multiplication.