remove references to local_rvs

Author: ferrine

pymc/variational/approximations.py

@@ -313,15 +313,7 @@ class SingleGroupApproximation(Approximation):
     _group_class = None
 
     def __init__(self, *args, **kwargs):
-        local_rv = kwargs.get("local_rv")
         groups = [self._group_class(None, *args, **kwargs)]
-        if local_rv is not None:
-            groups.extend(
-                [
-                    Group([v], params=p, local=True, model=kwargs.get("model"))
-                    for v, p in local_rv.items()
-                ]
-            )
         super().__init__(groups, model=kwargs.get("model"))
 
     def __getattr__(self, item):
@@ -360,8 +352,6 @@ class Empirical(SingleGroupApproximation):
     _group_class = EmpiricalGroup
 
     def __init__(self, trace=None, size=None, **kwargs):
-        if kwargs.get("local_rv", None) is not None:
-            raise opvi.LocalGroupError("Empirical approximation does not support local variables")
         super().__init__(trace=trace, size=size, **kwargs)
 
     def evaluate_over_trace(self, node):

pymc/variational/inference.py

@@ -22,7 +22,7 @@
 
 import pymc as pm
 
-from pymc.variational import opvi, test_functions
+from pymc.variational import test_functions
 from pymc.variational.approximations import Empirical, FullRank, MeanField
 from pymc.variational.operators import KL, KSD
 
@@ -334,7 +334,7 @@ class ADVI(KLqp):
     The last ones are local random variables
     :math:`{\cal Z}=\{\mathbf{z}_{i}\}_{i=1}^{N}`, where
     :math:`\mathbf{z}_{i}=\{\mathbf{z}_{i}^{k}\}_{k=1}^{V_{l}}`.
-    These RVs are used only in AEVB.
+    These RVs are used only in AEVB (which is not implemented in PyMC).
 
     The goal of ADVI is to approximate the posterior distribution
     :math:`p(\Theta,{\cal Z}|{\cal Y})` by variational posterior
@@ -408,8 +408,8 @@ class ADVI(KLqp):
 
     -   The probabilistic model
 
-        `model` with three types of RVs (`observed_RVs`,
-        `global_RVs` and `local_RVs`).
+        `model` with two types of RVs (`observed_RVs`,
+        `global_RVs`).
 
     -   (optional) Minibatches
 
@@ -428,10 +428,6 @@ class ADVI(KLqp):
 
     Parameters
     ----------
-    local_rv: dict[var->tuple]
-        mapping {model_variable -> approx params}
-        Local Vars are used for Autoencoding Variational Bayes
-        See (AEVB; Kingma and Welling, 2014) for details
     model: :class:`pymc.Model`
         PyMC model for inference
     random_seed: None or int
@@ -463,10 +459,6 @@ class FullRankADVI(KLqp):
 
     Parameters
     ----------
-    local_rv: dict[var->tuple]
-        mapping {model_variable -> approx params}
-        Local Vars are used for Autoencoding Variational Bayes
-        See (AEVB; Kingma and Welling, 2014) for details
     model: :class:`pymc.Model`
         PyMC model for inference
     random_seed: None or int
@@ -571,8 +563,6 @@ def __init__(
         kernel=test_functions.rbf,
         **kwargs,
     ):
-        if kwargs.get("local_rv") is not None:
-            raise opvi.AEVBInferenceError("SVGD does not support local groups")
         empirical = Empirical(
             size=n_particles,
             jitter=jitter,
@@ -639,9 +629,7 @@ def __init__(self, approx=None, estimator=KSD, kernel=test_functions.rbf, **kwar
             "is often **underestimated** when using temperature = 1."
         )
         if approx is None:
-            approx = FullRank(
-                model=kwargs.pop("model", None), local_rv=kwargs.pop("local_rv", None)
-            )
+            approx = FullRank(model=kwargs.pop("model", None))
         super().__init__(estimator=estimator, approx=approx, kernel=kernel, **kwargs)
 
     def fit(

pymc/variational/opvi.py

@@ -1145,9 +1145,6 @@ class Approximation(WithMemoization):
         -   :class:`FullRank`
         -   :class:`Empirical`
 
-    Single group accepts `local_rv` keyword with dict mapping PyMC variables
-    to their local Group parameters dict
-
     See Also
     --------
     :class:`Group`