Mixing precision, i.e. data types double and float etc., requires a precise handling of type casting. In #33 with the commit c51ecc1 a proposal has been given.
This behavior is not included in the NSL::Tensor class, see #57, which has to be updated once we agree on a scheme.
evanberkowitz
commented
2 years ago
Mixing precision, i.e. data types double and float etc., requires a precise handling of type casting. In #33 with the commit c51ecc1 a proposal has been given.
This behavior is not included in the
NSL::Tensorclass, see #57, which has to be updated once we agree on a scheme.