What is the most effective method for learning the NXM matrix (for each sample) in relation to different size input vectors and matrices?
To make it more clear:
I have a 56x32 output matrix for my first sample, and the corresponding input set is five different vectors and matrices, with their sizes as follows: 1x18, 1x18, 96x18, 96x18, and 1x56
Their size is always the same for every sample, but I am completely lost as to how to arrange the input and output. How can I arrange these different-sized vectors and matrices to learn a matrix output?
The number of samples will be at least 1000, maybe more if it is possible.
How can I learn about a settlement, or more importantly, can I learn at all? How do you solve this problem?
(Ps: The values are mostly float numbers)
Best
