# Copyright (c) 2017-2019 Uber Technologies, Inc.
# SPDX-License-Identifier: Apache-2.0
import torch
from torch.distributions import constraints
from torch.distributions.transforms import Transform
from pyro.distributions.util import copy_docs_from
[docs]@copy_docs_from(Transform)
class LowerCholeskyAffine(Transform):
"""
A bijection of the form,
:math:`\\mathbf{y} = \\mathbf{L} \\mathbf{x} + \\mathbf{r}`
where `\\mathbf{L}` is a lower triangular matrix and `\\mathbf{r}` is a vector.
:param loc: the fixed D-dimensional vector to shift the input by.
:type loc: torch.tensor
:param scale_tril: the D x D lower triangular matrix used in the transformation.
:type scale_tril: torch.tensor
"""
codomain = constraints.real_vector
bijective = True
event_dim = 1
volume_preserving = False
def __init__(self, loc, scale_tril):
super().__init__(cache_size=1)
self.loc = loc
self.scale_tril = scale_tril
assert loc.size(-1) == scale_tril.size(-1) == scale_tril.size(-2), \
"loc and scale_tril must be of size D and D x D, respectively (instead: {}, {})".format(loc.shape,
scale_tril.shape)
def _call(self, x):
"""
:param x: the input into the bijection
:type x: torch.Tensor
Invokes the bijection x=>y; in the prototypical context of a
:class:`~pyro.distributions.TransformedDistribution` `x` is a sample from
the base distribution (or the output of a previous transform)
"""
return torch.matmul(self.scale_tril, x.unsqueeze(-1)).squeeze(-1) + self.loc
def _inverse(self, y):
"""
:param y: the output of the bijection
:type y: torch.Tensor
Inverts y => x.
"""
return torch.triangular_solve((y - self.loc).unsqueeze(-1), self.scale_tril,
upper=False, transpose=False)[0].squeeze(-1)
[docs] def log_abs_det_jacobian(self, x, y):
"""
Calculates the elementwise determinant of the log Jacobian, i.e.
log(abs(dy/dx)).
"""
return torch.ones(x.size()[:-1], dtype=x.dtype, layout=x.layout, device=x.device) * \
self.scale_tril.diag().log().sum()