The current parallel pool does not have SPMD enabled. To use multi-GPU or parallel training, use 'delete(gcp)' to close the current pool and open a new one with SPMD enabled


trainNetwork