andreaferretti
commented
7 years ago I think this would be a great addition, but as you can see from the TODO, there are many other things that would be nice... :-)
If you are willing to contribute, though, I think it would make much sense to have sparse analogues of the core types, with sparse BLAS operations and functions to convert them to dense ones when needed
Hello Andrea, Great work on this library.
For large sparse datasets and for machine learning having sparse matrix support would be great to minimize memory usage.
It might be better to have that in a completely different library, in that case feel free to close the ticket.