Moving data would be a bottleneck for sure. Distributing the model itself with the state of the calculation and the required samples is just too much compared t

Moving data would be a bottleneck for sure. Distributing the model itself with the state of the calculation and the required samples is just too much compared to the CPU power a raspberry pi can provide.