Stochastic node in pymc v5

Hello everyone,
I’ve been studying shared random streams in order to include a variable in my model that’s random but not updated. This discussion makes sense to me, but I can’t seem to implement something similar in v5.

import pymc as pm
import numpy as np
import pytensor.tensor as pt

srng = pt.random.utils.RandomStream(seed=123)
true_mu, true_sigma = 5., 2.
y_obs = np.random.randn(50) * true_sigma + true_mu

with pm.Model() as m:
    mu = pm.Deterministic('mu', srng.normal(4.9, .1))
    sigma = pm.HalfCauchy('sigma', 5.)
    y = pm.Normal('y', mu, sigma, observed=y_obs)
    trace = pm.sample()

gives me

ValueError: Random variables detected in the logp graph: {normal_rv{0, (0, 0), floatX, False}.out}.
This can happen when DensityDist logp or Interval transform functions reference nonlocal variables,
or when not all rvs have a corresponding value variable.

Any ideas? Thank you very much!

This is explicitly not allowed. If you want to do it you have to implement your own step sampler that just returns draws from the prior. To my knowledge there is about 0 work out there suggesting you would ever want to do this.