Is there a way to store multiple matrices in a vector in octave?

Question

I am quite new to octave and have the following question: In Mathematica, you can initialize a vector, say t = {0,0,0} and can use each entry of t to store a matrix of possibly different size. For instance you could set

t[[1]] = IdentityMatrix[3];
t[[2]] = IdentityMatrix[4];
t[[3]] = IdentityMatrix[5];

giving you a 'tensor' with different dimensions for different first indices. Is there any equivalent way in Octave? My naive attempts failed. I would like to use it in a neural network, where I need to reshape a long vector into several matrices depending on the number and size of the layers, which I would like to store in a vector for convenience of scalability.

Is this or something similar possible?

Thanks in advance!

Answer 1

This is what cell arrays are for.

E.g.

t = { eye(3), eye(4), eye(5) }

gives

t =
{
  [1,1] =
Diagonal Matrix
     1   0   0
     0   1   0
     0   0   1
  [1,2] =
Diagonal Matrix
     1   0   0   0
     0   1   0   0
     0   0   1   0
     0   0   0   1
  [1,3] =
Diagonal Matrix
     1   0   0   0   0
     0   1   0   0   0
     0   0   1   0   0
     0   0   0   1   0
     0   0   0   0   1
}