torch.distributions.transforms.StickBreakingTransform Class Reference

Inheritance diagram for torch.distributions.transforms.StickBreakingTransform:

Public Member Functions
def	__eq__ (self, other)
def	log_abs_det_jacobian (self, x, y)
Public Member Functions inherited from torch.distributions.transforms.Transform
def	__init__ (self, cache_size=0)
def	inv (self)
def	sign (self)
def	__eq__ (self, other)
def	__ne__ (self, other)
def	__call__ (self, x)
def	log_abs_det_jacobian (self, x, y)
def	__repr__ (self)

Static Public Attributes
	domain
	codomain
	bijective
	event_dim
Static Public Attributes inherited from torch.distributions.transforms.Transform
	bijective
	event_dim

Detailed Description

Transform from unconstrained space to the simplex of one additional
dimension via a stick-breaking process.

This transform arises as an iterated sigmoid transform in a stick-breaking
construction of the `Dirichlet` distribution: the first logit is
transformed via sigmoid to the first probability and the probability of
everything else, and then the process recurses.

This is bijective and appropriate for use in HMC; however it mixes
coordinates together and is less appropriate for optimization.

Definition at line 473 of file transforms.py.

---

The documentation for this class was generated from the following file:

• torch/distributions/transforms.py