Enforce custom initval in SMC

pymc/smc/kernels.py

@@ -53,8 +53,9 @@ class SMC_KERNEL(ABC):
         initialize_population
             Choose initial population of SMC particles. Should return a dictionary
             with {var.name : numpy array of size (draws, var.size)}. Defaults
-            to sampling from the prior distribution. This method is only called
-            if `start` is not specified.
+            to sampling from the prior distribution, except for parameters which have custom
+            `initval`, in which case that value is used for all SMC particles.
+            This method is only called if `start` is not specified.
 
         _initialize_kernel : default
             Creates initial population of particles in the variable
@@ -144,7 +145,8 @@ def __init__(
             independent chains. Defaults to 2000.
         start : dict, or array of dict, default None
             Starting point in parameter space. It should be a list of dict with length `chains`.
-            When None (default) the starting point is sampled from the prior distribution.
+            When None (default) the starting point is sampled from the prior distribution, except
+            for parameters with a custom `initval`, in which case that value is used.
         model : Model (optional if in ``with`` context).
         random_seed : int, array_like of int, RandomState or Generator, optional
             Value used to initialize the random number generator.
@@ -193,10 +195,13 @@ def initialize_population(self) -> dict[str, np.ndarray]:
             )
 
             model = self.model
+
             prior_expression = make_initial_point_expression(
                 free_rvs=model.free_RVs,
                 rvs_to_transforms=model.rvs_to_transforms,
-                initval_strategies={},
+                initval_strategies={
+                    **model.rvs_to_initial_values,
+                },
                 default_strategy="prior",
                 return_transformed=True,
             )

tests/smc/test_smc.py

@@ -25,6 +25,7 @@
 import pymc as pm
 
 from pymc.backends.base import MultiTrace
+from pymc.distributions.transforms import Ordered
 from pymc.pytensorf import floatX
 from pymc.smc.kernels import IMH, systematic_resampling
 from tests.helpers import assert_random_state_equal
@@ -269,6 +270,30 @@ def test_deprecated_abc_args(self):
             ):
                 pm.sample_smc(draws=10, chains=1, save_log_pseudolikelihood=True)
 
+    def test_ordered(self):
+        """
+        Test that initial population respects custom initval, especially when applied
+        to the Ordered transformation. Regression test for #7438.
+        """
+        with pm.Model() as m:
+            pm.Normal(
+                "a",
+                mu=0.0,
+                sigma=1.0,
+                size=(2,),
+                transform=Ordered(),
+                initval=[-1.0, 1.0],
+            )
+
+        smc = IMH(model=m)
+        out = smc.initialize_population()
+
+        # initial point should not include NaNs
+        assert not np.any(np.isnan(out["a_ordered__"]))
+
+        # initial point should match for all particles
+        assert np.all(out["a_ordered__"][0] == out["a_ordered__"])
+
 
 class TestMHKernel:
     def test_normal_model(self):