by contrast if I use N=10000. I get a few warnings - but it does finish and plots something sensible
Auto-assigning NUTS sampler...
Initializing NUTS using jitter+adapt_diag...
/opt/conda/lib/python3.10/site-packages/pymc/aesaraf.py:1005: UserWarning: The parameter 'updates' of aesara.function() expects an OrderedDict, got <class 'dict'>. Using a standard dictionary here results in non-deterministic behavior. You should use an OrderedDict if you are using Python 2.7 (collections.OrderedDict for older python), or use a list of (shared, update) pairs. Do not just convert your dictionary to this type before the call as the conversion will still be non-deterministic.
aesara_function = aesara.function(
Sequential sampling (2 chains in 1 job)
NUTS: [mu, sigma]
100.00% [2000/2000 00:27<00:00 Sampling chain 0, 0 divergences]
/opt/conda/lib/python3.10/site-packages/aesara/scalar/basic.py:3070: RuntimeWarning: divide by zero encountered in log1p
return np.log1p(x)
/opt/conda/lib/python3.10/site-packages/aesara/scalar/basic.py:2001: RuntimeWarning: divide by zero encountered in divide
return x / y
100.00% [2000/2000 00:26<00:00 Sampling chain 1, 0 divergences]
/opt/conda/lib/python3.10/site-packages/aesara/scalar/basic.py:3070: RuntimeWarning: divide by zero encountered in log1p
return np.log1p(x)
/opt/conda/lib/python3.10/site-packages/aesara/scalar/basic.py:2001: RuntimeWarning: divide by zero encountered in divide
return x / y
Sampling 2 chains for 1_000 tune and 1_000 draw iterations (2_000 + 2_000 draws total) took 54 seconds.
We recommend running at least 4 chains for robust computation of convergence diagnostics
100.00% [2000/2000 00:01<00:00]