def add_param(self, user=None, name=None, ref=0.0, dtype="floatX"):
if dtype == "floatX":
dtype = aesara.config.floatX
spec = self.__param_spec__[name]
shape = tuple(eval(s, {"d": self.dim}) for s in spec)
if user is None:
if self.local:
raise opvi.LocalGroupError(
"Need parameters for local group flow")
if self.batched:
if self.batch_size is None:
raise opvi.BatchedGroupError(
"Need batch size to infer parameter shape")
shape = (self.batch_size, ) + shape
return aesara.shared(
np.asarray(np.random.normal(size=shape) * self.__jitter +
ref).astype(dtype),
name=name,
)
else:
if self.batched:
if self.local or self.batch_size is None:
shape = (-1, ) + shape
else:
shape = (self.batch_size, ) + shape
return at.as_tensor(user).reshape(shape)
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)