# Copyright Contributors to the Pyro project.
# SPDX-License-Identifier: Apache-2.0
import functools
import math
import operator
import weakref
from typing import Dict
import numpy as np
import torch
from numpy.polynomial.hermite import hermgauss
class _SafeLog(torch.autograd.Function):
@staticmethod
def forward(ctx, x):
ctx.save_for_backward(x)
return x.log()
@staticmethod
def backward(ctx, grad):
(x,) = ctx.saved_tensors
return grad / x.clamp(min=torch.finfo(x.dtype).eps)
[docs]def safe_log(x):
"""
Like :func:`torch.log` but avoids infinite gradients at log(0)
by clamping them to at most ``1 / finfo.eps``.
"""
return _SafeLog.apply(x)
[docs]def log_beta(x, y, tol=0.0):
"""
Computes log Beta function.
When ``tol >= 0.02`` this uses a shifted Stirling's approximation to the
log Beta function. The approximation adapts Stirling's approximation of the
log Gamma function::
lgamma(z) ≈ (z - 1/2) * log(z) - z + log(2 * pi) / 2
to approximate the log Beta function::
log_beta(x, y) ≈ ((x-1/2) * log(x) + (y-1/2) * log(y)
- (x+y-1/2) * log(x+y) + log(2*pi)/2)
The approximation additionally improves accuracy near zero by iteratively
shifting the log Gamma approximation using the recursion::
lgamma(x) = lgamma(x + 1) - log(x)
If this recursion is applied ``n`` times, then absolute error is bounded by
``error < 0.082 / n < tol``, thus we choose ``n`` based on the user
provided ``tol``.
:param torch.Tensor x: A positive tensor.
:param torch.Tensor y: A positive tensor.
:param float tol: Bound on maximum absolute error. Defaults to 0.1. For
very small ``tol``, this function simply defers to :func:`log_beta`.
:rtype: torch.Tensor
"""
assert isinstance(tol, (float, int)) and tol >= 0
if tol < 0.02:
# At small tolerance it is cheaper to defer to torch.lgamma().
return x.lgamma() + y.lgamma() - (x + y).lgamma()
# This bound holds for arbitrary x,y. We could do better with large x,y.
shift = int(math.ceil(0.082 / tol))
xy = x + y
factors = []
for _ in range(shift):
factors.append(xy / (x * y))
x = x + 1
y = y + 1
xy = xy + 1
log_factor = functools.reduce(operator.mul, factors).log()
return (
log_factor
+ (x - 0.5) * x.log()
+ (y - 0.5) * y.log()
- (xy - 0.5) * xy.log()
+ (math.log(2 * math.pi) / 2 - shift)
)
[docs]@torch.no_grad()
def log_binomial(n, k, tol=0.0):
"""
Computes log binomial coefficient.
When ``tol >= 0.02`` this uses a shifted Stirling's approximation to the
log Beta function via :func:`log_beta`.
:param torch.Tensor n: A nonnegative integer tensor.
:param torch.Tensor k: An integer tensor ranging in ``[0, n]``.
:rtype: torch.Tensor
"""
assert isinstance(tol, (float, int)) and tol >= 0
n_plus_1 = n + 1
if tol < 0.02:
# At small tolerance it is cheaper to defer to torch.lgamma().
return n_plus_1.lgamma() - (k + 1).lgamma() - (n_plus_1 - k).lgamma()
return -n_plus_1.log() - log_beta(k + 1, n_plus_1 - k, tol=tol)
[docs]def log_I1(orders: int, value: torch.Tensor, terms=250):
r"""Compute first n log modified bessel function of first kind
.. math ::
\log(I_v(z)) = v*\log(z/2) + \log(\sum_{k=0}^\inf \exp\left[2*k*\log(z/2) - \sum_kk^k log(kk)
- \lgamma(v + k + 1)\right])
:param orders: orders of the log modified bessel function.
:param value: values to compute modified bessel function for
:param terms: truncation of summation
:return: 0 to orders modified bessel function
"""
orders = orders + 1
if len(value.size()) == 0:
vshape = torch.Size([1])
else:
vshape = value.shape
value = value.reshape(-1, 1)
k = torch.arange(terms, device=value.device)
lgammas_all = torch.lgamma(torch.arange(1, terms + orders + 1, device=value.device))
assert lgammas_all.shape == (orders + terms,) # lgamma(0) = inf => start from 1
lvalues = torch.log(value / 2) * k.view(1, -1)
assert lvalues.shape == (vshape.numel(), terms)
lfactorials = lgammas_all[:terms]
assert lfactorials.shape == (terms,)
lgammas = lgammas_all.repeat(orders).view(orders, -1)
assert lgammas.shape == (orders, terms + orders) # lgamma(0) = inf => start from 1
indices = k[:orders].view(-1, 1) + k.view(1, -1)
assert indices.shape == (orders, terms)
seqs = (
2 * lvalues[None, :, :]
- lfactorials[None, None, :]
- lgammas.gather(1, indices)[:, None, :]
).logsumexp(-1)
assert seqs.shape == (orders, vshape.numel())
i1s = lvalues[..., :orders].T + seqs
assert i1s.shape == (orders, vshape.numel())
return i1s.view(-1, *vshape)
[docs]def get_quad_rule(num_quad, prototype_tensor):
r"""
Get quadrature points and corresponding log weights for a Gauss Hermite quadrature rule
with the specified number of quadrature points.
Example usage::
quad_points, log_weights = get_quad_rule(32, prototype_tensor)
# transform to N(0, 4.0) Normal distribution
quad_points *= 4.0
# compute variance integral in log-space using logsumexp and exponentiate
variance = torch.logsumexp(quad_points.pow(2.0).log() + log_weights, axis=0).exp()
assert (variance - 16.0).abs().item() < 1.0e-6
:param int num_quad: number of quadrature points.
:param torch.Tensor prototype_tensor: used to determine `dtype` and `device` of returned tensors.
:return: tuple of `torch.Tensor`s of the form `(quad_points, log_weights)`
"""
quad_rule = hermgauss(num_quad)
quad_points = quad_rule[0] * np.sqrt(2.0)
log_weights = np.log(quad_rule[1]) - 0.5 * np.log(np.pi)
return torch.from_numpy(quad_points).type_as(prototype_tensor), torch.from_numpy(
log_weights
).type_as(prototype_tensor)
[docs]def sparse_multinomial_likelihood(total_count, nonzero_logits, nonzero_value):
"""
The following are equivalent::
# Version 1. dense
log_prob = Multinomial(logits=logits).log_prob(value).sum()
# Version 2. sparse
nnz = value.nonzero(as_tuple=True)
log_prob = sparse_multinomial_likelihood(
value.sum(-1),
(logits - logits.logsumexp(-1))[nnz],
value[nnz],
)
"""
return (
_log_factorial_sum(total_count)
- _log_factorial_sum(nonzero_value)
+ torch.dot(nonzero_logits, nonzero_value)
)
_log_factorial_cache: Dict[int, torch.Tensor] = {}
def _log_factorial_sum(x: torch.Tensor) -> torch.Tensor:
if x.requires_grad:
return (x + 1).lgamma().sum()
key = id(x)
if key not in _log_factorial_cache:
weakref.finalize(x, _log_factorial_cache.pop, key, None) # type: ignore
_log_factorial_cache[key] = (x + 1).lgamma().sum()
return _log_factorial_cache[key]