Dimension error for Hierarchical model

Questions version agnostic
7

You can rely on numpy broadcasting semantics:

import numpy as np

x = np.zeros(120)
y = np.zeros(5)
z1 = x[:, None] + y
z2 = x + y[:, None]

assert z1.shape == (120, 5)
assert z2.shape == (5, 120)

Works the same with PyMC variables