+ &jB(Rt^RIHtHt^RIHt.ROt!RR4t]!4t]!4t ]P]P4] P]P4R44t ]P]P4R4t] P]P4R4t]P]P 4]P]P"4] P]P 4] P]P"4R4444t]P]P&4]P]P(4] P]P&4] P]P(4R 4444t]P]P,4] P]P,4R 44t]P]P04]P]P24] P]P04] P]P24R 4444t]P]P64R 4t] P]P64R 4t] P]P<4R4t] P]P@4] P]PB4R44t"]P]PF4] P]PF4R44t$]P]PJ4R4t&] P]PJ4R4t']P]PP4R4t)] P]PP4R4t*R#)aF PyTorch provides two global :class:`ConstraintRegistry` objects that link :class:`~torch.distributions.constraints.Constraint` objects to :class:`~torch.distributions.transforms.Transform` objects. These objects both input constraints and return transforms, but they have different guarantees on bijectivity. 1. ``biject_to(constraint)`` looks up a bijective :class:`~torch.distributions.transforms.Transform` from ``constraints.real`` to the given ``constraint``. The returned transform is guaranteed to have ``.bijective = True`` and should implement ``.log_abs_det_jacobian()``. 2. ``transform_to(constraint)`` looks up a not-necessarily bijective :class:`~torch.distributions.transforms.Transform` from ``constraints.real`` to the given ``constraint``. The returned transform is not guaranteed to implement ``.log_abs_det_jacobian()``. The ``transform_to()`` registry is useful for performing unconstrained optimization on constrained parameters of probability distributions, which are indicated by each distribution's ``.arg_constraints`` dict. These transforms often overparameterize a space in order to avoid rotation; they are thus more suitable for coordinate-wise optimization algorithms like Adam:: loc = torch.zeros(100, requires_grad=True) unconstrained = torch.zeros(100, requires_grad=True) scale = transform_to(Normal.arg_constraints["scale"])(unconstrained) loss = -Normal(loc, scale).log_prob(data).sum() The ``biject_to()`` registry is useful for Hamiltonian Monte Carlo, where samples from a probability distribution with constrained ``.support`` are propagated in an unconstrained space, and algorithms are typically rotation invariant.:: dist = Exponential(rate) unconstrained = torch.zeros(100, requires_grad=True) sample = biject_to(dist.support)(unconstrained) potential_energy = -dist.log_prob(sample).sum() .. note:: An example where ``transform_to`` and ``biject_to`` differ is ``constraints.simplex``: ``transform_to(constraints.simplex)`` returns a :class:`~torch.distributions.transforms.SoftmaxTransform` that simply exponentiates and normalizes its inputs; this is a cheap and mostly coordinate-wise operation appropriate for algorithms like SVI. In contrast, ``biject_to(constraints.simplex)`` returns a :class:`~torch.distributions.transforms.StickBreakingTransform` that bijects its input down to a one-fewer-dimensional space; this a more expensive less numerically stable transform but is needed for algorithms like HMC. The ``biject_to`` and ``transform_to`` objects can be extended by user-defined constraints and transforms using their ``.register()`` method either as a function on singleton constraints:: transform_to.register(my_constraint, my_transform) or as a decorator on parameterized constraints:: @transform_to.register(MyConstraintClass) def my_factory(constraint): assert isinstance(constraint, MyConstraintClass) return MyTransform(constraint.param1, constraint.param2) You can create your own registry by creating a new :class:`ConstraintRegistry` object. ) constraints transforms)_NumberConstraintRegistrycFaa]tRt^PtoRtV3RltRRltRtRtVt V;t #)rz- Registry to link constraints to transforms. c2</Vn\SV` 4R#N) _registrysuper__init__)self __class__s&/Users/jameslopez/projects/CWCArchive/cwc-podcast/.venv/lib/python3.14/site-packages/torch/distributions/constraint_registry.pyr ConstraintRegistry.__init__Us c aaVfVV3Rl#\S\P4'd \S4o\S\4'd!\ S\P4'g\ RS 24hVSP S&V#)a Registers a :class:`~torch.distributions.constraints.Constraint` subclass in this registry. Usage:: @my_registry.register(MyConstraintClass) def construct_transform(constraint): assert isinstance(constraint, MyConstraint) return MyTransform(constraint.arg_constraints) Args: constraint (subclass of :class:`~torch.distributions.constraints.Constraint`): A subclass of :class:`~torch.distributions.constraints.Constraint`, or a singleton object of the desired class. factory (Callable): A callable that inputs a constraint object and returns a :class:`~torch.distributions.transforms.Transform` object. c(<SPSV4#r)register)factory constraintr s&r-ConstraintRegistry.register..ls4==W#ErzLExpected constraint to be either a Constraint subclass or instance, but got ) isinstancer Constrainttype issubclass TypeErrorr r rrsff&rrConstraintRegistry.registerYs$ ?E E j+"8"8 9 9j)J*d++:  ..4 4 ^_i^jk &-z"rc VP\V4,pT!T4# \d%\R\T4P R24Rhi;i)a Looks up a transform to constrained space, given a constraint object. Usage:: constraint = Normal.arg_constraints["scale"] scale = transform_to(constraint)(torch.zeros(1)) # constrained u = transform_to(constraint).inv(scale) # unconstrained Args: constraint (:class:`~torch.distributions.constraints.Constraint`): A constraint object. Returns: A :class:`~torch.distributions.transforms.Transform` object. Raises: `NotImplementedError` if no transform has been registered. zCannot transform z constraintsN)r rKeyErrorNotImplementedError__name__rs&& r__call__ConstraintRegistry.__call__|s`( nnT*%56G z""  %#D$4$=$=#>lK  s &/A)r r) r" __module__ __qualname____firstlineno____doc__r rr#__static_attributes____classdictcell__ __classcell__)r __classdict__s@@rrrPs!F##rc"\P#r)ridentity_transformrs&r_transform_to_realr0s  ( ((rcl\VP4p\P!WP4#r) biject_tobase_constraintrIndependentTransformreinterpreted_batch_ndimsrbase_transforms& r_biject_to_independentr8s.z99:N  * *<< rcl\VP4p\P!WP4#r) transform_tor3rr4r5r6s& r_transform_to_independentr;s.!*"<"<=N  * *<< rc,\P!4#r)r ExpTransformr/s&r_transform_to_positiver>s  " " $$rc\P!\P!4\P!VP^4.4#))rComposeTransformr=AffineTransform lower_boundr/s&r_transform_to_greater_thanrDs>  & &  # # %  & &z'='=q A  rc\P!\P!4\P!VPR4.4#)r@)rrAr=rB upper_boundr/s&r_transform_to_less_thanrHs>  & &  # # %  & &z'='=r B  rc\VP\4;'dVP^8Hp\VP\4;'dVP^8HpV'dV'd\P !4#VPpVPVP, p\P !\P !4\P!W44.4#))rrCrrGrSigmoidTransformrArB)r lower_is_0 upper_is_1locscales& r_transform_to_intervalrPs :))73SS 8N8NRS8S :))73SS 8N8NRS8Sj**,,  C  " "Z%;%; ;E  & &  $ $ & (B(B3(NO rc,\P!4#r)rStickBreakingTransformr/s&r_biject_to_simplexrS  , , ..rc,\P!4#r)rSoftmaxTransformr/s&r_transform_to_simplexrWs  & & ((rc,\P!4#r)rLowerCholeskyTransformr/s&r_transform_to_lower_choleskyrZrTrc,\P!4#r)rPositiveDefiniteTransformr/s&r_transform_to_positive_definiter]s  / / 11rc,\P!4#r)rCorrCholeskyTransformr/s&r_transform_to_corr_choleskyr`s  + + --rc\P!VPUu.uFp\V4NK upVPVP 4#uupir)r CatTransformcseqr2dimlengthsrcs& r_biject_to_catrh s@  " ")/!1/ASAS /Ac\P!VPUu.uFp\V4NK upVPVP 4#uupir)rrbrcr:rdrerfs& r_transform_to_catrks@  " "",//2/Qa/2JNNJDVDV 2ric\P!VPUu.uFp\V4NK upVP4#uupir)rStackTransformrcr2rdrfs& r_biject_to_stackrns8  $ $)/!1/ /Ac\P!VPUu.uFp\V4NK upVP4#uupir)rrmrcr:rdrfs& r_transform_to_stackrqs8  $ $",//2/Qa/2JNN 2roN)rr2r:)+r(torch.distributionsrr torch.typesr__all__rr2r:rrealr0 independentr8r;positive nonnegativer> greater_thangreater_than_eqrD less_thanrHintervalhalf_open_intervalrPsimplexrSrWlower_choleskyrZpositive_definitepositive_semidefiniter] corr_choleskyr`catrhrkstackrnrqrrrsAF8 F#F#R  !#  K$$% {''())&) K++,-{../0 K(() K++, {++, {../%0--*% K,,- K//0 {//0 {223411. K))* {,,-.+ K(() K223 {++, {5567-4*$ K''(/)/{**+),) {112/3/{445 {8892:62 K--. {001.2/. KOO$% {'(  K%%&' {(()*r