Source code for pyro.poutine.replay_messenger

from .messenger import Messenger

[docs]class ReplayMessenger(Messenger): """ Given a callable that contains Pyro primitive calls, return a callable that runs the original, reusing the values at sites in trace at those sites in the new trace Consider the following Pyro program: >>> def model(x): ... s = pyro.param("s", torch.tensor(0.5)) ... z = pyro.sample("z", dist.Normal(x, s)) ... return z ** 2 ``replay`` makes ``sample`` statements behave as if they had sampled the values at the corresponding sites in the trace: >>> old_trace = pyro.poutine.trace(model).get_trace(1.0) >>> replayed_model = pyro.poutine.replay(model, trace=old_trace) >>> bool(replayed_model(0.0) == old_trace.nodes["_RETURN"]["value"]) True :param fn: a stochastic function (callable containing Pyro primitive calls) :param trace: a :class:`~pyro.poutine.Trace` data structure to replay against :param params: dict of names of param sites and constrained values in fn to replay against :returns: a stochastic function decorated with a :class:`~pyro.poutine.replay_messenger.ReplayMessenger` """ def __init__(self, trace=None, params=None): """ :param trace: a trace whose values should be reused Constructor. Stores trace in an attribute. """ super(ReplayMessenger, self).__init__() if trace is None and params is None: raise ValueError("must provide trace or params to replay against") self.trace = trace self.params = params def _pyro_sample(self, msg): """ :param msg: current message at a trace site. At a sample site that appears in self.trace, returns the value from self.trace instead of sampling from the stochastic function at the site. At a sample site that does not appear in self.trace, reverts to default Messenger._pyro_sample behavior with no additional side effects. """ name = msg["name"] if self.trace is not None and name in self.trace: guide_msg = self.trace.nodes[name] if msg["is_observed"]: return None if guide_msg["type"] != "sample" or \ guide_msg["is_observed"]: raise RuntimeError("site {} must be sample in trace".format(name)) msg["done"] = True msg["value"] = guide_msg["value"] msg["infer"] = guide_msg["infer"] return None def _pyro_param(self, msg): name = msg["name"] if self.params is not None and name in self.params: assert hasattr(self.params[name], "unconstrained"), \ "param {} must be constrained value".format(name) msg["done"] = True msg["value"] = self.params[name] return None