Error autogenerating process inputs when data batch size does not match Simulator.minibatch_size

[drasmuss]

import nengo
import numpy as np

import nengo_dl

with nengo.Network() as net:
    a = nengo.Node([0])
    b = nengo.Node(nengo.processes.WhiteNoise())

with nengo_dl.Simulator(net, minibatch_size=2) as sim:
    sim.predict(np.zeros((10, 1, 1)))

Gives error

ValueError: cannot copy sequence with size 2 to array axis with dimension 10

This is because with processes we generate a number of instances equal to the minibatch_size. But then when we call predict with a batch size of 10 (for the first node) but no inputs for the second node, we need to autogenerate data with a batch size of 10 for the second node (even though we only have 2 instances).