PR #76 made me remember that I generally dislike the data container approach. I really feel like the shared model approach is generally better. However, it is clumsy to use and leads to boilerplate code. In Mango.jl we implemented an approach to reduce the overhead. This issue shall track the transfer of this feature to mango python.
For the feature description see the original issue OFFIS-DAI/Mango.jl#24.
PR #76 made me remember that I generally dislike the data container approach. I really feel like the shared model approach is generally better. However, it is clumsy to use and leads to boilerplate code. In Mango.jl we implemented an approach to reduce the overhead. This issue shall track the transfer of this feature to mango python.
For the feature description see the original issue OFFIS-DAI/Mango.jl#24.