Given two tensors A and B with the same dimension (d>=2)
and shapes [A_{1},...,A_{d-2},A_{d-1},A_{d}]
and [A_{1},...,A_{d-2},B_{d-1},B_{d}]
(shapes of the first d-2 dimensions are identical).
Is there a way to calculate the kronecker product over the last two dimensions?
The shape of my_kron(A,B)
should be [A_{1},...,A_{d-2},A_{d-1}*B_{d-1},A_{d}*B_{d}]
.
For example with d=3
,
A.shape=[2,3,3]
B.shape=[2,4,4]
C=my_kron(A,B)
C[0,...]
should be the kronecker product of A[0,...]
and B[0,...]
and C[1,...]
the kronecker product of A[1,...]
and B[1,...]
.
For d=2 this is simply what the jnp.kron
(or np.kron
) function does.
For d=3 this can be achived with jax.vmap
.
jax.vmap(lambda x, y: jnp.kron(x[0, :], y[0, :]))(A, B)
But I was not able to find a solution for general (unknown) dimensions. Any suggestions?
In
numpy
terms I think this is what you are doing:That treats the initial dimension, the 2, as a
batch
. One obvious generalization is to reshape the arrays, reducing the higher dimensions to 1, e.g.reshape(-1,3,3)
, etc. And then afterwards, reshapeC
back to the desired n-dimensions.np.kron
does accept 3d (and higher), but it's doing some sort ofouter
on the shared 2 dimension:And visualizing that 4 dimension as (2,2), I can take the
diagonal
and get yourC
:The full
kron
does more calculations than needed, but is still faster:I'm sure it's possible to do your calculation in a more direct way, but doing that requires a better understanding of how the
kron
works. A quick glance as thenp.kron
code suggest that is does anouter(A,B)
which has the same number of elements, but it then
reshapes
andconcatenates
to produce thekron
layout.But following a hunch, I found that this is equivalent to what you want:
That is easily generalized to more leading dimensions.