3
O4W^1  ã               @   sT   d dl mZ d dlZd dlmZ d dlmZ d dlmZ dd„ Z	G dd	„ d	eƒZ
dS )
é    )ÚNumberN)Úconstraints)ÚExponentialFamily)Úbroadcast_allc             C   s
   t j| ƒS )N)ÚtorchÚ_standard_gamma)Úconcentration© r	   ú:/tmp/pip-build-e66j97b2/torch/torch/distributions/gamma.pyr   	   s    r   c                   s˜   e Zd ZdZejejdœZejZdZdZ	e
dd„ ƒZe
dd„ ƒZd‡ fd
d„	Zd‡ fdd„	Zejƒ fdd„Zdd„ Zdd„ Ze
dd„ ƒZdd„ Z‡  ZS )ÚGammaa	  
    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)
    )r   ÚrateTr   c             C   s   | j | j S )N)r   r   )Úselfr	   r	   r
   Úmean"   s    z
Gamma.meanc             C   s   | j | jjdƒ S )Né   )r   r   Úpow)r   r	   r	   r
   Úvariance&   s    zGamma.varianceNc                sR   t ||ƒ\| _| _t|tƒr0t|tƒr0tjƒ }n
| jjƒ }tt	| ƒj
||d� d S )N)Úvalidate_args)r   r   r   Ú
isinstancer   r   ÚSizeÚsizeÚsuperr   Ú__init__)r   r   r   r   Úbatch_shape)Ú	__class__r	   r
   r   *   s
    

zGamma.__init__c                sR   | j t|ƒ}tj|ƒ}| jj|ƒ|_| jj|ƒ|_tt|ƒj|dd� | j	|_	|S )NF)r   )
Z_get_checked_instancer   r   r   r   Úexpandr   r   r   Ú_validate_args)r   r   Z	_instanceÚnew)r   r	   r
   r   2   s    
zGamma.expandc             C   sD   | j |ƒ}t| jj|ƒƒ| jj|ƒ }|jƒ jtj|j	ƒj
d� |S )N)Úmin)Z_extended_shaper   r   r   r   ÚdetachZclamp_r   ZfinfoÚdtypeZtiny)r   Zsample_shapeÚshapeÚvaluer	   r	   r
   Úrsample;   s    
zGamma.rsamplec             C   sd   t j|| jj| jjd�}| jr(| j|ƒ | jt j| jƒ | jd t j|ƒ  | j|  t j	| jƒ S )N)r   Údeviceé   )
r   Z	as_tensorr   r   r#   r   Z_validate_sampler   ÚlogÚlgamma)r   r!   r	   r	   r
   Úlog_probA   s    
zGamma.log_probc             C   s4   | j tj| jƒ tj| j ƒ d| j  tj| j ƒ  S )Ng      ð?)r   r   r%   r   r&   Zdigamma)r   r	   r	   r
   ÚentropyI   s    zGamma.entropyc             C   s   | j d | j fS )Nr$   )r   r   )r   r	   r	   r
   Ú_natural_paramsM   s    zGamma._natural_paramsc             C   s&   t j|d ƒ|d t j|jƒ  ƒ  S )Nr$   )r   r&   r%   Z
reciprocal)r   ÚxÚyr	   r	   r
   Ú_log_normalizerQ   s    zGamma._log_normalizer)N)N)Ú__name__Ú
__module__Ú__qualname__Ú__doc__r   ZpositiveZarg_constraintsZsupportZhas_rsampleZ_mean_carrier_measureÚpropertyr   r   r   r   r   r   r"   r'   r(   r)   r,   Ú__classcell__r	   r	   )r   r
   r      s   	r   )Znumbersr   r   Ztorch.distributionsr   Ztorch.distributions.exp_familyr   Ztorch.distributions.utilsr   r   r   r	   r	   r	   r
   Ú<module>   s   