pytorch: Is there a way to maintain and multiply a unit vector matrix in GPU?
I have a giant sparse matrix S. Each column has 1 entry whose value is 1.
torch: Is there a way to maintain and multiply a unit vector matrix in gpu
I have a giant sparse matrix S. Each column has 1 entry whose value is 1. The rest of the values are zeros. I frequently have to multiply S with another matrix A (S@A). It’s really just a “selector”, so it’s trivial to implement in numpy.