About the usage of sparse_add

[lebron-2016]

Dear author,

Thanks for your great work!

I am a little confused about the correct use of sparse_add, when I use a network with SparseInverseConv. If I want to use Fsp.sparse_add(a,b) correctly, do I only need to ensure that the one with more activation values ​​in a and b has indice_dict and its activation position contains all the activation positions of the other one?

Looking forward to your reply.

Thanks a lot!!