mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-06 12:20:52 +01:00
173f224570
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/18598 ghimport-source-id: c74597e5e7437e94a43c163cee0639b20d0d0c6a Stack from [ghstack](https://github.com/ezyang/ghstack): * **#18598 Turn on F401: Unused import warning.** This was requested by someone at Facebook; this lint is turned on for Facebook by default. "Sure, why not." I had to noqa a number of imports in __init__. Hypothetically we're supposed to use __all__ in this case, but I was too lazy to fix it. Left for future work. Be careful! flake8-2 and flake8-3 behave differently with respect to import resolution for # type: comments. flake8-3 will report an import unused; flake8-2 will not. For now, I just noqa'd all these sites. All the changes were done by hand. Signed-off-by: Edward Z. Yang <ezyang@fb.com> Differential Revision: D14687478 fbshipit-source-id: 30d532381e914091aadfa0d2a5a89404819663e3
82 lines
3.0 KiB
Python
82 lines
3.0 KiB
Python
from numbers import Number
|
|
|
|
import torch
|
|
from torch.distributions import constraints
|
|
from torch.distributions.exp_family import ExponentialFamily
|
|
from torch.distributions.utils import broadcast_all
|
|
|
|
|
|
def _standard_gamma(concentration):
|
|
return torch._standard_gamma(concentration)
|
|
|
|
|
|
class Gamma(ExponentialFamily):
|
|
r"""
|
|
Creates a Gamma distribution parameterized by shape :attr:`concentration` and :attr:`rate`.
|
|
|
|
Example::
|
|
|
|
>>> m = Gamma(torch.tensor([1.0]), torch.tensor([1.0]))
|
|
>>> m.sample() # Gamma distributed with concentration=1 and rate=1
|
|
tensor([ 0.1046])
|
|
|
|
Args:
|
|
concentration (float or Tensor): shape parameter of the distribution
|
|
(often referred to as alpha)
|
|
rate (float or Tensor): rate = 1 / scale of the distribution
|
|
(often referred to as beta)
|
|
"""
|
|
arg_constraints = {'concentration': constraints.positive, 'rate': constraints.positive}
|
|
support = constraints.positive
|
|
has_rsample = True
|
|
_mean_carrier_measure = 0
|
|
|
|
@property
|
|
def mean(self):
|
|
return self.concentration / self.rate
|
|
|
|
@property
|
|
def variance(self):
|
|
return self.concentration / self.rate.pow(2)
|
|
|
|
def __init__(self, concentration, rate, validate_args=None):
|
|
self.concentration, self.rate = broadcast_all(concentration, rate)
|
|
if isinstance(concentration, Number) and isinstance(rate, Number):
|
|
batch_shape = torch.Size()
|
|
else:
|
|
batch_shape = self.concentration.size()
|
|
super(Gamma, self).__init__(batch_shape, validate_args=validate_args)
|
|
|
|
def expand(self, batch_shape, _instance=None):
|
|
new = self._get_checked_instance(Gamma, _instance)
|
|
batch_shape = torch.Size(batch_shape)
|
|
new.concentration = self.concentration.expand(batch_shape)
|
|
new.rate = self.rate.expand(batch_shape)
|
|
super(Gamma, new).__init__(batch_shape, validate_args=False)
|
|
new._validate_args = self._validate_args
|
|
return new
|
|
|
|
def rsample(self, sample_shape=torch.Size()):
|
|
shape = self._extended_shape(sample_shape)
|
|
value = _standard_gamma(self.concentration.expand(shape)) / self.rate.expand(shape)
|
|
value.detach().clamp_(min=torch.finfo(value.dtype).tiny) # do not record in autograd graph
|
|
return value
|
|
|
|
def log_prob(self, value):
|
|
if self._validate_args:
|
|
self._validate_sample(value)
|
|
return (self.concentration * torch.log(self.rate) +
|
|
(self.concentration - 1) * torch.log(value) -
|
|
self.rate * value - torch.lgamma(self.concentration))
|
|
|
|
def entropy(self):
|
|
return (self.concentration - torch.log(self.rate) + torch.lgamma(self.concentration) +
|
|
(1.0 - self.concentration) * torch.digamma(self.concentration))
|
|
|
|
@property
|
|
def _natural_params(self):
|
|
return (self.concentration - 1, -self.rate)
|
|
|
|
def _log_normalizer(self, x, y):
|
|
return torch.lgamma(x + 1) + (x + 1) * torch.log(-y.reciprocal())
|