clipppy.guide.sampling_group

Module Contents

clipppy.guide.sampling_group._Tensor_Type

class clipppy.guide.sampling_group.LocatedAndScaledSamplingGroupWithPrior(sites: Iterable[clipppy.utils.typing._Site], name='', *args, **kwargs)

class clipppy.guide.sampling_group.LocatedSamplingGroup(sites: Iterable[clipppy.utils.typing._Site], name='', *args, **kwargs)

loc :torch.Tensor

loc(self)

class clipppy.guide.sampling_group.SamplingGroup(sites: Iterable[clipppy.utils.typing._Site], name='', *args, **kwargs)

include_det_jac = True

_cat_sites(self, vals: Mapping[str, _Tensor_Type]) → _Tensor_Type

_process_prototype(self)

abstract _sample(self, infer: dict = None) → torch.Tensor

_sample_site(self, group_z: torch.Tensor, name: str, fn: pyro.distributions.TorchDistribution = None)

static _scale_diagonal(scale: Union[torch.Tensor, float], jac: torch.Tensor)

static _scale_matrix(scale: Union[torch.Tensor, float], jac: torch.Tensor)

property event_shape(self) → torch.Size

extra_repr(self) → str

Set the extra representation of the module

To print customized extra information, you should re-implement this method in your own modules. Both single-line and multi-line strings are acceptable.

forward(self, infer: dict = None) → tuple[torch.Tensor, MutableMapping[str, torch.Tensor]]

property grad_context(self)

init(self) → torch.Tensor

jacobian(self, guide_z: torch.Tensor, sites: Iterable[str] = None) → torch.Tensor

mask(self) → torch.BoolTensor

unpack(self, group_z: torch.Tensor, sites: Mapping[str, clipppy.utils.typing._Site] = None, guiding=True) → MutableMapping[str, torch.Tensor]

unpack_site(self, arr: torch.Tensor, name: str)

class clipppy.guide.sampling_group.SamplingGroupWithPrior(sites: Iterable[clipppy.utils.typing._Site], name='', *args, **kwargs)

guide_z :torch.Tensor

_sample(self, infer=None) → torch.Tensor

guide_z(self)

abstract prior(self)

class clipppy.guide.sampling_group.ScaledSamplingGroup(sites, name='', init_scale: Union[torch.Tensor, float] = 1.0, *args, **kwargs)