ISubGVQA.sampling.methods.noise
===============================

.. py:module:: ISubGVQA.sampling.methods.noise


Attributes
----------

.. autoapisummary::

   ISubGVQA.sampling.methods.noise.logger


Classes
-------

.. autoapisummary::

   ISubGVQA.sampling.methods.noise.BaseNoiseDistribution
   ISubGVQA.sampling.methods.noise.SumOfGammaNoiseDistribution
   ISubGVQA.sampling.methods.noise.GumbelDistribution


Module Contents
---------------

.. py:data:: logger

.. py:class:: BaseNoiseDistribution

   Bases: :py:obj:`abc.ABC`


   Helper class that provides a standard way to create an ABC using
   inheritance.


   .. py:method:: sample(shape: torch.Size) -> torch.Tensor
      :abstractmethod:



.. py:class:: SumOfGammaNoiseDistribution(k: float, nb_iterations: int = 10, device: Optional[torch.device] = None)

   Bases: :py:obj:`BaseNoiseDistribution`


   Creates a generator of samples for the Sum-of-Gamma distribution [1], parameterized
   by :attr:`k`, :attr:`nb_iterations`, and :attr:`device`.

   [1] Mathias Niepert, Pasquale Minervini, Luca Franceschi - Implicit MLE: Backpropagating Through Discrete
   Exponential Family Distributions. NeurIPS 2021 (https://arxiv.org/abs/2106.01798)

   Example::

       >>> import torch
       >>> noise_distribution = SumOfGammaNoiseDistribution(k=5, nb_iterations=100)
       >>> noise_distribution.sample(torch.Size([5]))
       tensor([ 0.2504,  0.0112,  0.5466,  0.0051, -0.1497])

   Args:
       k (float): k parameter -- see [1] for more details.
       nb_iterations (int): number of iterations for estimating the sample.
       device (torch.devicde): device where to store samples.


   .. py:attribute:: k


   .. py:attribute:: nb_iterations
      :value: 10



   .. py:attribute:: device
      :value: None



   .. py:method:: sample(shape: torch.Size) -> torch.Tensor


.. py:class:: GumbelDistribution(loc: float = 0.0, scale: float = 1.0, device: torch.device = 'cpu')

   Bases: :py:obj:`BaseNoiseDistribution`


   Helper class that provides a standard way to create an ABC using
   inheritance.


   .. py:attribute:: loc
      :value: 0.0



   .. py:attribute:: _scale
      :value: 1.0



   .. py:attribute:: device
      :value: 'cpu'



   .. py:property:: scale


   .. py:method:: sample(shape: torch.Size) -> torch.Tensor