
    [Th                         S SK r S SK Jr  S SKJr  S SKJr  S SKJr  S SKJ	r	  S SK
JrJrJrJrJr  S SKJrJr  S	S
/r " S S	\5      r " S S
\5      rg)    N)Tensor)constraints)Distribution)TransformedDistribution)SigmoidTransform)broadcast_allclamp_probslazy_propertylogits_to_probsprobs_to_logits)_Number_sizeLogitRelaxedBernoulliRelaxedBernoullic                   (  ^  \ rS rSrSr\R                  \R                  S.r\R                  r	SU 4S jjr
SU 4S jjrS r\S\4S j5       r\S\4S	 j5       r\S\R&                  4S
 j5       r\R&                  " 5       4S\S\4S jjrS rSrU =r$ )r      aO  
Creates a LogitRelaxedBernoulli distribution parameterized by :attr:`probs`
or :attr:`logits` (but not both), which is the logit of a RelaxedBernoulli
distribution.

Samples are logits of values in (0, 1). See [1] for more details.

Args:
    temperature (Tensor): relaxation temperature
    probs (Number, Tensor): the probability of sampling `1`
    logits (Number, Tensor): the log-odds of sampling `1`

[1] The Concrete Distribution: A Continuous Relaxation of Discrete Random
Variables (Maddison et al., 2017)

[2] Categorical Reparametrization with Gumbel-Softmax
(Jang et al., 2017)
probslogitsc                   > Xl         US L US L :X  a  [        S5      eUb#  [        U[        5      n[	        U5      u  U l        O"[        U[        5      n[	        U5      u  U l        Ub  U R
                  OU R                  U l        U(       a  [        R                  " 5       nOU R                  R                  5       n[        TU ]1  XdS9  g )Nz;Either `probs` or `logits` must be specified, but not both.validate_args)temperature
ValueError
isinstancer   r   r   r   _paramtorchSizesizesuper__init__)selfr   r   r   r   	is_scalarbatch_shape	__class__s          ]/var/www/auris/envauris/lib/python3.13/site-packages/torch/distributions/relaxed_bernoulli.pyr!   LogitRelaxedBernoulli.__init__,   s    &TMv~.M  "5'2I)%0MTZ"673I*62NT[$)$5djj4;;**,K++**,KB    c                   > U R                  [        U5      n[        R                  " U5      nU R                  Ul        SU R
                  ;   a1  U R                  R                  U5      Ul        UR                  Ul        SU R
                  ;   a1  U R                  R                  U5      Ul	        UR                  Ul        [        [        U]/  USS9  U R                  Ul        U$ )Nr   r   Fr   )_get_checked_instancer   r   r   r   __dict__r   expandr   r   r    r!   _validate_argsr"   r$   	_instancenewr%   s       r&   r,   LogitRelaxedBernoulli.expand?   s    (()>	Jjj-**dmm#

))+6CICJt}}$++K8CJCJ#S2;e2T!00
r(   c                 :    U R                   R                  " U0 UD6$ N)r   r0   )r"   argskwargss      r&   _newLogitRelaxedBernoulli._newM   s    {{///r(   returnc                 *    [        U R                  SS9$ NT)	is_binary)r   r   r"   s    r&   r   LogitRelaxedBernoulli.logitsP   s    tzzT::r(   c                 *    [        U R                  SS9$ r:   )r   r   r<   s    r&   r   LogitRelaxedBernoulli.probsT   s    t{{d;;r(   c                 6    U R                   R                  5       $ r3   )r   r   r<   s    r&   param_shape!LogitRelaxedBernoulli.param_shapeX   s    {{!!r(   sample_shapec                 t   U R                  U5      n[        U R                  R                  U5      5      n[        [        R
                  " X#R                  UR                  S95      nUR                  5       U* R                  5       -
  UR                  5       -   U* R                  5       -
  U R                  -  $ )N)dtypedevice)_extended_shaper	   r   r,   r   randrE   rF   loglog1pr   )r"   rC   shaper   uniformss        r&   rsampleLogitRelaxedBernoulli.rsample\   s    $$\2DJJ--e45JJuKKE
 LLNxi..00599;>5&AQQ 	r(   c                 .   U R                   (       a  U R                  U5        [        U R                  U5      u  p!X!R	                  U R
                  5      -
  nU R
                  R                  5       U-   SUR                  5       R                  5       -  -
  $ )N   )	r-   _validate_sampler   r   mulr   rI   exprJ   )r"   valuer   diffs       r&   log_probLogitRelaxedBernoulli.log_probf   su    !!%(%dkk59		$"2"233##%,q488:3C3C3E/EEEr(   )r   r   r   r   NNNr3   )__name__
__module____qualname____firstlineno____doc__r   unit_intervalrealarg_constraintssupportr!   r,   r6   r
   r   r   r   propertyr   r   rA   r   rM   rV   __static_attributes____classcell__r%   s   @r&   r   r      s    & !, 9 9[EUEUVOGC&0 ; ; ; <v < < "UZZ " " -2JJL E V F Fr(   c                      ^  \ rS rSrSr\R                  \R                  S.r\R                  r	Sr
SU 4S jjrSU 4S jjr\S\4S j5       r\S\4S	 j5       r\S\4S
 j5       rSrU =r$ )r   n   a  
Creates a RelaxedBernoulli distribution, parametrized by
:attr:`temperature`, and either :attr:`probs` or :attr:`logits`
(but not both). This is a relaxed version of the `Bernoulli` distribution,
so the values are in (0, 1), and has reparametrizable samples.

Example::

    >>> # xdoctest: +IGNORE_WANT("non-deterministic")
    >>> m = RelaxedBernoulli(torch.tensor([2.2]),
    ...                      torch.tensor([0.1, 0.2, 0.3, 0.99]))
    >>> m.sample()
    tensor([ 0.2951,  0.3442,  0.8918,  0.9021])

Args:
    temperature (Tensor): relaxation temperature
    probs (Number, Tensor): the probability of sampling `1`
    logits (Number, Tensor): the log-odds of sampling `1`
r   Tc                 L   > [        XU5      n[        TU ]	  U[        5       US9  g )Nr   )r   r    r!   r   )r"   r   r   r   r   	base_distr%   s         r&   r!   RelaxedBernoulli.__init__   s'    )+fE	$4$6mTr(   c                 J   > U R                  [        U5      n[        TU ]  XS9$ )N)r/   )r*   r   r    r,   r.   s       r&   r,   RelaxedBernoulli.expand   s'    (()99Ew~k~99r(   r8   c                 .    U R                   R                  $ r3   )ri   r   r<   s    r&   r   RelaxedBernoulli.temperature   s    ~~)))r(   c                 .    U R                   R                  $ r3   )ri   r   r<   s    r&   r   RelaxedBernoulli.logits   s    ~~$$$r(   c                 .    U R                   R                  $ r3   )ri   r   r<   s    r&   r   RelaxedBernoulli.probs   s    ~~###r(    rX   r3   )rY   rZ   r[   r\   r]   r   r^   r_   r`   ra   has_rsampler!   r,   rb   r   r   r   r   rc   rd   re   s   @r&   r   r   n   s    ( !, 9 9[EUEUVO''GKU: *V * * % % % $v $ $r(   )r   r   torch.distributionsr    torch.distributions.distributionr   ,torch.distributions.transformed_distributionr   torch.distributions.transformsr   torch.distributions.utilsr   r	   r
   r   r   torch.typesr   r   __all__r   r   rs   r(   r&   <module>r|      sQ      + 9 P ;  ' #$6
7VFL VFr+$. +$r(   