+
    ýäûj�@  ã                   ó  € ^ RI HtHtHtHtHtHtHtHtH	t	 ^ RI
t
^ RI
Ht ^ RIHtHt ^ RIHtHt ^ RIHtHt ]! R4      t]! 4       tR R	 lt ! R
 R]4      t ! R R]P6                  ]R7      t ! R R]4      t ! R R]]4      tR# )é    )	ÚAnyÚCallableÚDictÚOptionalÚSetÚTupleÚTypeÚTypeVarÚUnionN)Únn)ÚDistributionÚUniform)ÚMultiprocessWrapperÚ_holds_graph_tensor)Ú_get_rebuild_deviceÚ_set_rebuild_deviceÚTc                óF   € V ^8„  d   QhRRR\         \        R3,          RR/# )é   Ú	generatorÚRandomGeneratorBaseÚargs.ÚreturnN)r   r   )Úformats   "ÚeC:\Ai_Coding\apps\studio\ComfyUI\.venv\Lib\site-packages\kornia\augmentation\random_generator\base.pyÚ__annotate__r       s3   € ÷ 	Gñ 	GÐ-Bð 	GÌ%ÔPSÐUXÐPXÍ/ð 	GÐ^bñ 	Gó    c                 óB  € V P                  4       '       d‰   \        V P                  4      pV P                  V P                  V P
                  4       \        V P                  4      V,
           F+  pV P                  P                  V4      V P                  V&   K-  	  R # R # ©N)Ú_ranges_need_rebuildÚsetÚ_parametersÚ_rebuild_samplersÚdeviceÚdtypeÚpopÚ__dict__)r   r   Ú
registeredÚnames   &&  r   Ú_rebuild_samplers_before_drawr*       s€   € à×%Ñ%×'Ò'Ü˜×.Ñ.Ó/ˆ
Ø×#Ñ# I×$4Ñ$4°i·o±oÔFô ˜	×-Ñ-Ó.°×;Ð;ˆDØ'0×'<Ñ'<×'@Ñ'@ÀÓ'FˆI×Ñ˜tÓ$ó <ñ (r   c                   ó6   a € ] tR t^,t o RtV 3R lR ltRtV tR# )Ú_PostInitInjectionMetaClasszMTo inject the ``__post_init__`` function after the creation of each instance.c                óB   <€ V ^8„  d   QhRS[ S[,          RS[RS[RS[/# )r   Úclsr   Úkwargsr   )r	   r   r   )r   Ú__classdict__s   "€r   r   Ú(_PostInitInjectionMetaClass.__annotate__/   s-   ø€ ÷ ñ ‘d™1•gð ¡cð ±Sð ¹Qñ r   c                óZ   € \         P                  ! V .VO5/ VB pVP                  4        V# r   )ÚtypeÚ__call__Ú__post_init__)r.   r   r/   Úobjs   &*, r   r4   Ú$_PostInitInjectionMetaClass.__call__/   s+   € Ü�mŠm˜CÐ1 $Ò1¨&Ñ1ˆØ×ÑÔØˆ
r   © N)Ú__name__Ú
__module__Ú__qualname__Ú__firstlineno__Ú__doc__r4   Ú__static_attributes__Ú__classdictcell__©r0   s   @r   r,   r,   ,   s   ø‡ € ÙW÷ö r   r,   c                   ó¢  a a€ ] tR t^5t oRtRtRtV3R lV 3R lltV3R lR ltR]	P                  3V3R lR	 lltV3R
 lR ltV3R lR ltV3R lR ltV3R lR ltV3R lR ltV3R lV 3R lltV3R lR ltV3R lV 3R lltV3R lV 3R lltV3R lV 3R lltV3R lV 3R lltV3R  lR! ltR&V3R" lR# lltV3R$ ltR%tVtV ;t# )'r   z9Base class for generating random augmentation parameters.NTc                ó   <€ V ^8„  d   QhRR/# ©r   r   Nr8   )r   r0   s   "€r   r   Ú RandomGeneratorBase.__annotate__@   s   ø€ ÷ ñ ˜$ñ r   c                ó$   <€ \         SV `  4        R # r   )ÚsuperÚ__init__)ÚselfÚ	__class__s   &€r   rG   ÚRandomGeneratorBase.__init__@   s   ø€ Ü‰ÑÖr   c                ó   <€ V ^8„  d   QhRR/# rC   r8   )r   r0   s   "€r   r   rD   C   s   ø€ ÷ (ñ (˜tñ (r   c                ó&   € V P                  4        R # r   )Úset_rng_device_and_dtype©rH   s   &r   r5   Ú!RandomGeneratorBase.__post_init__C   s   € Ø×%Ñ%Ö'r   c                ób   <€ V ^8„  d   QhRS[ S[P                  ,          RS[P                  RR/# ©r   r$   r%   r   N)r   Útorchr$   r%   )r   r0   s   "€r   r   rD   F   s6   ø€ ÷ ,ñ ,á™Ÿ™Õ&ð,ñ �{‰{ð,ð 
ñ	,r   c           
     óx  € Vf   \         P                  ! R4      pV P                  V8w  g   V P                  V8w  dø   \        V P                  4      pV P                  4       '       d   V P                  W4       MV P                  W4       V P                  P                  4        UUu0 uF#  w  rEVP                  V\        4      VJg   K!  VkK%  	  ppp\        \        VP                  V P                  P                  RR4      4      4      4      V n        Wn        W n        V P                  4        R# R# u uppi )z’Change the random generation device and dtype.

Note:
    The generated random numbers are not reproducible across different devices and dtypes.

NÚcpuÚ_sampler_attrsr8   )rR   r$   r%   Údictr'   Ú_follows_rangesr#   Úmake_samplersÚitemsÚgetÚ_UNSETÚtupleÚsortedÚunionrU   Ú_follow_learnable_ranges)rH   r$   r%   Úbeforer)   ÚvalueÚmades   &&&    r   rM   Ú,RandomGeneratorBase.set_rng_device_and_dtypeF   sô   € ð Š>Ü—\’\ %Ó(ˆFØ�;‰;˜&Ô  D§J¡J°%Ô$7Ü˜$Ÿ-™-Ó(ˆFØ×#Ñ#×%Ò%Ø×&Ñ& vÕ5à×"Ñ" 6Ô1ð -1¯M©M×,?Ñ,?Ô,AÔkÑ,A™[˜TÀVÇZÁZÐPTÔV\ÓE]ÐejÐEj—D�DÑ,AˆDÑkÜ"'¬¨t¯z©z¸$¿-¹-×:KÑ:KÐL\Ð^`Ó:aÓ/bÓ(cÓ"dˆDÔØ ŒKØŒJØ×)Ñ)Ö+ñ %8ùó ls   Â(D6ÃD6c                ój   <€ V ^8„  d   QhRS[ S[S[P                  R3,          RS[P                  RR/# )r   r$   Nr%   r   )r   ÚstrrR   r$   r%   )r   r0   s   "€r   r   rD   a   s7   ø€ ÷ 'ñ '©©c±5·<±<ÀÐ.EÕ(Fð 'ÉuÏ{É{ð 'Ð_cñ 'r   c                ó²   € \        4       p\        \        P                  ! V4      4        V P	                  W4       \        V4       R #   \        T4       i ; ir   )r   r   rR   r$   rX   )rH   r$   r%   Úouters   &&& r   r#   Ú%RandomGeneratorBase._rebuild_samplersa   sB   € ô $Ó%ˆÜœEŸLšL¨Ó0Ô1ð	'Ø×Ñ˜vÔ-ä Ö&øÔ Õ&ús   «A	 Á	Ac                ó    <€ V ^8„  d   QhRS[ /# ©r   r   ©Úbool)r   r0   s   "€r   r   rD   k   s   ø€ ÷ Qñ Q¡ñ Qr   c                óD   € \         V P                  P                  4       9   # r   )r*   Ú_forward_pre_hooksÚvaluesrN   s   &r   rW   Ú#RandomGeneratorBase._follows_rangesk   s   € Ü,°×0GÑ0G×0NÑ0NÓ0PÑPÐPr   c                ó   <€ V ^8„  d   QhRR/# rC   r8   )r   r0   s   "€r   r   rD   n   s   ø€ ÷ Jñ J¨$ñ Jr   c                óf  a € S P                   '       d   S P                  4       '       d   R # S P                  P                  RR4      pS P                  '       gA   \
        ;QJ d    V 3R lV 4       F  '       g   K   RM	  RM! V 3R lV 4       4      '       d   S P                  \        4       R # R # )NrU   c              3   ól   <"  € T F)  p\        SP                  P                  V4      4      x € K+  	  R # 5ir   )r   r'   rZ   ©Ú.0r)   rH   s   & €r   Ú	<genexpr>Ú?RandomGeneratorBase._follow_learnable_ranges.<locals>.<genexpr>u   s,   øé € Ð"jÑ\iÐTXÔ#6°t·}±}×7HÑ7HÈÓ7N×#OÐ#OÓ\iùs   ƒ14TFr8   )Ú_follows_ranges_in_forwardrW   r'   rZ   r"   ÚanyÚregister_forward_pre_hookr*   )rH   Úsampler_attrss   f r   r_   Ú,RandomGeneratorBase._follow_learnable_rangesn   s|   ø€ ð ×.×.Ð.°$×2FÑ2F×2HÒ2HÙØŸ™×)Ñ)Ð*:¸BÓ?ˆØ××ÐŸs›sÔ"jÑ\iÓ"jŸsŸsšsÔ"jÑ\iÓ"j×jÒjØ×*Ñ*Ô+HÖIñ  kr   c                ó>   <€ V ^8„  d   QhRS[ RS[P                  RR/# )r   r)   ra   r   N)re   rR   ÚTensor)r   r0   s   "€r   r   rD   x   s&   ø€ ÷ #ñ #©#ð #±e·l±lð #Àtñ #r   c                ó–   € \        V\        P                  4      '       g   V P                  P	                  VR 4       \        WV4       R # r   )Ú
isinstancer   Ú	Parameterr"   r&   Úsetattr)rH   r)   ra   s   &&&r   Ú_assign_derived_rangeÚ)RandomGeneratorBase._assign_derived_rangex   s5   € ô ˜%¤§¡×.Ò.Ø×Ñ× Ñ   tÔ,Ü�˜EÖ"r   c                ó    <€ V ^8„  d   QhRS[ /# rj   rk   )r   r0   s   "€r   r   rD   €   s   ø€ ÷ ñ ¡dñ r   c                ó  a€ \         P                  ! V P                  4      P                  oV P                  P	                  4        Uu. uF
  qf   K  VNK  	  pp\
        ;QJ d    V3R lV 4       F  '       d   K   RM	  RM! V3R lV 4       4      '       d   R# \         P                  ! 4       \        ;QJ d    . R V 4       F  NK  	  5M! R V 4       4      3pW0P                  P                  R4      8X  d   R# W0n
        R# u upi )Nc              3   óT   <"  € T F  qP                   P                  S8H  x € K  	  R # 5ir   )r$   r3   )ru   ra   Údevice_types   & €r   rv   Ú;RandomGeneratorBase._ranges_need_rebuild.<locals>.<genexpr>ˆ   s   øé € ÐD¹V°E�|‰|× Ñ  KÖ/»Vùs   ƒ%(FTc              3   óP   "  € T F  p\        V4      VP                  3x € K  	  R # 5ir   )ÚidÚ_version)ru   ra   s   & r   rv   r‰   Š   s    é € Ð2cÑ\bÐSX´B°u³I¸u¿~¹~Õ3NÓ\bùs   ‚$&Ú_range_versions)rR   r$   r3   r"   ro   ÚallÚis_grad_enabledr\   r'   rZ   r�   )rH   ra   ÚrangesÚversionsrˆ   s   &   @r   r    Ú(RandomGeneratorBase._ranges_need_rebuild€   s¾   ø€ ô —l’l 4§;¡;Ó/×4Ñ4ˆØ%)×%5Ñ%5×%<Ñ%<Ô%>ÓTÑ%>˜E—%�%Ñ%>ˆÐTß‹3ÔD¹VÓD�3�3Š3ÔD¹VÓD×DÒDÙÜ×)Ò)Ó+¯U¬UÑ2cÑ\bÓ2c¯U©UÑ2cÑ\bÓ2cÓ-cÐdˆØ—}‘}×(Ñ(Ð):Ó;Ô;ÙØ'ÔÙùò Us   ÁDÁDc          
      óL   <€ V ^8„  d   QhRS[ S[S[3,          RS[RS[RS[RR/# )r   Ú
state_dictÚprefixr   r/   r   N©r   re   r   )r   r0   s   "€r   r   rD   �   s;   ø€ ÷  ñ  ±±S¹#°Xµð  Éð  ÑTWð  Ñcfð  Ðkoñ  r   c           	     óX  <€ V P                    Uu0 uF  qRV,           V9   g   K  VkK  	  ppV P                  P                  4        Fä  w  rWVP                  W%,           4      p\	        V\
        P                  4      '       g   K>  \	        V\
        P                  4      '       g   K`  WPP                  9  g   Kr  VP                  VP                  8X  g   K�  \
        P                  ! VP                  VP                  VP                  R 7      V4      '       d   KÓ  VP                  V4       Kæ  	  V P                  P                  4        UUu/ uF-  w  rWWV9   g   K  WWVP                  4       P                  4       3bK/  	  p	pp\         SV `D  ! W.VO5/ VB  V'       g   R# \%        V P&                  4      p
 V P)                  V4       R# u upi u uppi   \*         dÌ    T	P                  4        Fa  w  pw  rxYpP                  T&   \
        P,                  ! 4       ;_uu_ 4        TP/                  T4       RRR4       KO    + '       g   i     K`  ; i	  \1        T P&                  4      \1        T
4      ,
           F  pT P&                  T K  	  T P&                  P3                  T
4       h i ; i)©r$   r%   N)r"   Ú_buffersrY   rZ   r€   rR   r~   Ú_non_persistent_buffers_setÚshapeÚequalÚtor$   r%   ÚaddÚdetachÚclonerF   Ú_load_from_state_dictrV   r'   Ú_ranges_loadedÚ	ExceptionÚno_gradÚcopy_r!   Úupdate)rH   r”   r•   r   r/   r)   Úchangedra   Úsavedr`   ÚstaterI   s   &&&*,      €r   r¡   Ú)RandomGeneratorBase._load_from_state_dict�   sô  ø€ ð %)×$4Ò$4ÓTÑ$4˜DÀ½ÈÑ8S—4�4Ñ$4ˆÐTØŸ=™=×.Ñ.Ö0‰KˆDØ—N‘N 6¥=Ó1ˆEä˜5¤%§,¡,×/Ô/Ü˜u¤e§l¡l×3Ô3Ø× @Ñ @Ö@Ø—K‘K 5§;¡;Ö.ÜŸš E§H¡H°E·L±LÈÏÉ HÓ$TÐV[×\Ô\à—‘˜DÖ!ñ 1ð LPÏ=É=×K^ÑK^ÔK`ÔtÑK`¹K¸DÐdhÑdsÔ7�$ §¡£× 4Ñ 4Ó 6Ð7Ò7ÑK`ˆÑtÜ‰Ò% jÐJ¸4ÒJÀ6ÒJßÙÜ�T—]‘]Ó#ˆð	Ø×Ñ Ö(ùò% Uùó uøô ô 
	ð )/¯©®Ñ$�‘n�uØ&+—‘˜dÑ#Ü—]’]—_•_Ø—K‘K Ô&÷ %—_—_’_úñ )7ô ˜DŸM™MÓ*¬S°«Z×7Ð7�Ø—M‘M $Ò'ñ 8à�M‰M× Ñ  Ô'Øð
	ús<   �G¤GÅ
GÅ$GÆ5G ÇAJ)È!H=	È3
J)È=IÉA!J)c                ó4   <€ V ^8„  d   QhRS[ S[,          RR/# )r   Únamesr   N)r   re   )r   r0   s   "€r   r   rD   ²   s   ø€ ÷ <ñ <¡C©¥Hð <°ñ <r   c                ó.  € V P                   P                  R R4       V P                  4       '       gO   \        P                  ! 4       ;_uu_ 4        V P                  V P                  V P                  4       RRR4       R# R#   + '       g   i     R# ; i)r�   N)r'   r&   rW   rR   Úenable_gradrX   r$   r%   )rH   r¬   s   &&r   r¢   Ú"RandomGeneratorBase._ranges_loaded²   sd   € à�‰×ÑÐ+¨TÔ2Ø×#Ñ#×%Ò%Ü×"Ò"×$Õ$Ø×"Ñ" 4§;¡;°·
±
Ô;÷ %Ñ$ñ &ß$×$Ð$ús   Á'BÂB	c                ó6   <€ V ^8„  d   QhRS[ S[S[3,          /# rj   r–   )r   r0   s   "€r   r   rD   ¹   s   ø€ ÷ ñ ™d¡3© 8�nñ r   c                ó$  <€ \         SV `  4       pVP                  R R4       Uu. uF'  p\        VP                  V4      4      '       g   K%  VNK)  	  ppV F  pW K  	  V'       d&   \        P
                  ! ^ V P                  R7      VR&   V# u upi )rU   )r$   Ú_sampler_device_prober8   )rF   Ú__getstate__rZ   r   rR   Úemptyr$   )rH   r©   r)   ÚdroppedrI   s   &   €r   r³   Ú RandomGeneratorBase.__getstate__¹   s„   ø€ ô
 ‘Ñ$Ó&ˆØ$)§I¡IÐ.>ÀÔ$CÓlÑ$C˜DÔGZÐ[`×[dÑ[dÐeiÓ[j×Gk—4�4Ñ$CˆÐlÛˆDØ’ñ çô .3¯[ª[¸À4Ç;Á;Ô-OˆEÐ)Ñ*Øˆùò ms   ¤"BÁBc                ó:   <€ V ^8„  d   QhRS[ S[S[3,          RR/# )r   r©   r   Nr–   )r   r0   s   "€r   r   rD   È   s#   ø€ ÷ (ñ (¡$¡s©C x¥.ð (°Tñ (r   c                ó<  <a € VP                  R R4      p\        SS `	  V4       \        ;QJ d9    V 3R lS P                  P                  RR4       4       F  '       g   K   RM.	  RM*! V 3R lS P                  P                  RR4       4       4      '       db   Ve   VP                  S n        \        P                  ! 4       ;_uu_ 4        S P                  S P                  S P                  4       RRR4       S P                  4        R#   + '       g   i     L"; i)r²   Nc              3   ó@   <"  € T F  qSP                   9  x € K  	  R # 5ir   )r'   rt   s   & €r   rv   Ú3RandomGeneratorBase.__setstate__.<locals>.<genexpr>Ë   s   øé € Ð]Ñ5\¨T˜4Ÿ=™=Ö(Ó5\ùs   ƒrU   TFr8   )r&   rF   Ú__setstate__ry   r'   rZ   r$   rR   r®   r#   r%   r_   )rH   r©   ÚproberI   s   f& €r   r»   Ú RandomGeneratorBase.__setstate__È   s¾   ù€ Ø—	‘	Ð1°4Ó8ˆÜ‰Ñ˜UÔ#ß‹3Ô]°T·]±]×5FÑ5FÐGWÐY[Ô5\Ó]�3�3Š3Ô]°T·]±]×5FÑ5FÐGWÐY[Ô5\Ó]×]Ò]ØÒ Ø#Ÿl™l�”ô ×"Ò"×$Õ$Ø×&Ñ& t§{¡{°D·J±JÔ?÷ %ð 	×%Ñ%Ö'÷ %×$ús   Ã
'DÄD	c                ó*   <€ V ^8„  d   QhRS[ RS[ RR/# )r   r   r/   r   r   )r   )r   r0   s   "€r   r   rD   Õ   s#   ø€ ÷ 
+ñ 
+™ð 
+¡sð 
+Ð/Dñ 
+r   c                ó$   <€ \         SV `  ! V/ VB # )zìUpdate sampler device and dtype using ``torch.nn.Module.to`` semantics.

Args:
    *args: Positional arguments accepted by ``Module.to``.
    **kwargs: Keyword arguments accepted by ``Module.to``.

Returns:
    This generator instance.
)rF   r�   )rH   r   r/   rI   s   &*,€r   r�   ÚRandomGeneratorBase.toÕ   s   ø€ ô ‰wŠz˜4Ð* 6Ñ*Ð*r   c                óp   <€ V ^8„  d   QhRS[ S[P                  .S[P                  3,          RS[RS[RR/# )r   Úfnr   r/   r   r   )r   rR   r~   r   )r   r0   s   "€r   r   rD   á   s?   ø€ ÷ ñ ™¡5§<¡< .±%·,±,Ð">Õ?ð Éð ÑWZð Ð_tñ r   c                ó*  <€ \         SV `  ! V.VO5/ VB pV! \        P                  ! RV P                  V P
                  R 7      4      pVP                  4       '       d   VP
                  MV P
                  pV P                  VP                  V4       V# )r˜   r8   )rF   Ú_applyrR   Úzerosr$   r%   Úis_floating_pointrM   )rH   rÂ   r   r/   Úoutr¼   r%   rI   s   &&*,   €r   rÄ   ÚRandomGeneratorBase._applyá   sr   ø€ ô ‰gŠn˜RÐ1 $Ò1¨&Ñ1ˆÙ”5—;’;˜r¨$¯+©+¸T¿Z¹ZÔHÓIˆØ$×6Ñ6×8Ò8�—’¸d¿j¹jˆØ×%Ñ% e§l¡l°EÔ:Øˆ
r   c                óR   <€ V ^8„  d   QhRS[ P                  RS[ P                  RR/# rQ   )rR   r$   r%   )r   r0   s   "€r   r   rD   ê   s*   ø€ ÷ "ñ "¡E§L¡Lð "¹¿¹ð "Èñ "r   c                ó   € \         h)aë  Create distribution samplers for the given device and dtype.

Assign the samplers to attributes of the generator, new or replaced, rather than writing them into an object
it already holds: a copy of the generator rebuilds the attributes this call assigns, and only those.

A generator that holds a learnable range goes through this again before each draw, with the checks in
:mod:`kornia.augmentation.utils.param_validation` off. If it reads the values of a range here anyway, set
``_follows_ranges_in_forward = False``: reading them would wait for the GPU and break a ``torch.compile`` graph.

Args:
    device: Target device.
    dtype: Target floating-point dtype.

Raises:
    NotImplementedError: Subclass did not implement sampler creation.
©ÚNotImplementedError)rH   r$   r%   s   &&&r   rX   Ú!RandomGeneratorBase.make_samplersê   s
   € ô" "Ð!r   c                ój   <€ V ^8„  d   QhRS[ S[R3,          RS[RS[S[S[P                  3,          /# )r   Úbatch_shape.Úsame_on_batchr   )r   Úintrl   r   re   rR   r~   )r   r0   s   "€r   r   rD   ý   s=   ø€ ÷ "ñ "¡5©¨c¨¥?ð "Á4ð "ÑTXÑY\Ñ^c×^jÑ^jÐYjÕTkñ "r   c                ó   € \         h)aN  Sample random augmentation parameters.

Args:
    batch_shape: Target batch shape.
    same_on_batch: If ``True``, use one sample and broadcast it across the
        batch dimension.

Returns:
    Dictionary of tensors consumed by augmentation modules.

Raises:
    NotImplementedError: Subclass did not implement parameter sampling.
rË   )rH   rÏ   rÐ   s   &&&r   ÚforwardÚRandomGeneratorBase.forwardý   s
   € ô "Ð!r   c                ó~   <€ V ^8„  d   Qh/ S[ S[S[P                  R3,          ;R&   S[P                  ;R&   S[;R&   # )r   Nr$   r%   rx   )r   re   rR   r$   r%   rl   )r   r0   s   "€r   r   rD   5   s@   ø‡ ‚ ñ ‘#‘u—|‘| TÐ)Õ*Ñ1ñ ñ �;‰;Ññ	 ñ !%Ñ+ò r   )r�   rU   r$   r%   )F)r9   r:   r;   r<   r=   r$   rx   rG   r5   rR   Úfloat32rM   r#   rW   r_   rƒ   r    r¡   r¢   r³   r»   r�   rÄ   rX   rÓ   Ú__annotate_func__r>   r?   Ú__classcell__)rI   r0   s   @@r   r   r   5   sÎ   ù‡ € ÙCà-1€Fð (,Ð÷ó ÷(ð (ð
 *.Ø"Ÿ]™]÷,ò ,÷6'ð '÷Qð Q÷Jð J÷#ð #÷ð ÷  ó  ÷D<ð <÷ó ÷(ó (÷
+ó 
+÷ó ÷"ð "÷&"ò "÷Q … r   r   )Ú	metaclassc                   ó‚   a € ] tR tRt o RtRV 3R lR lltV 3R lR ltV 3R lR	 ltV 3R
 lR ltV 3R lR lt	Rt
V tR# )ÚDistributionWithMapperi  a…  Wraps a distribution with a value mapper function.

This is used to restrict the output values of a given distribution by a value mapper function.
The value mapper function can be functions like sigmoid, tanh, etc.

Args:
    dist: the target distribution.
    map_fn: the callable function to adjust the output from distributions.

Example:
    >>> from torch.distributions import Normal
    >>> import torch.nn as nn
    >>> # without mapper
    >>> dist = DistributionWithMapper(Normal(0., 1.,), map_fn=None)
    >>> _ = torch.manual_seed(0)
    >>> dist.rsample((8,))
    tensor([ 1.5410, -0.2934, -2.1788,  0.5684, -1.0845, -1.3986,  0.4033,  0.8380])
    >>> # with sigmoid mapper
    >>> dist = DistributionWithMapper(Normal(0., 1.,), map_fn=nn.Sigmoid())
    >>> _ = torch.manual_seed(0)
    >>> dist.rsample((8,))
    tensor([0.8236, 0.4272, 0.1017, 0.6384, 0.2527, 0.1980, 0.5995, 0.6980])

Nc                óz   <€ V ^8„  d   QhRS[ RS[S[S[P                  .S[P                  3,          ,          RR/# )r   ÚdistÚmap_fnr   N)r   r   r   rR   r~   )r   r0   s   "€r   r   Ú#DistributionWithMapper.__annotate__(  s?   ø€ ÷ ñ ™\ð ±8¹HÁeÇlÁlÀ^ÑUZ×UaÑUaÐEaÕ<bÕ3cð Ðosñ r   c                ó   € Wn         W n        R # r   ©rÝ   rÞ   )rH   rÝ   rÞ   s   &&&r   rG   ÚDistributionWithMapper.__init__(  s   € ØŒ	ØŽr   c                óN   <€ V ^8„  d   QhRS[ S[R3,          RS[P                  /# ©r   Úsample_shape.r   ©r   rÑ   rR   r~   )r   r0   s   "€r   r   rß   ,  s'   ø€ ÷ ñ ¡E©#¨s¨(¥Oð ¹¿¹ñ r   c                ó¢   € V P                   P                  \        P                  ! V4      4      pV P                  e   V P	                  V4      pV# )z¥Draw a reparameterized sample and apply ``map_fn`` when provided.

Args:
    sample_shape: Desired sample shape.

Returns:
    Sample tensor after optional mapping.
)rÝ   ÚrsamplerR   ÚSizerÞ   ©rH   rå   rÇ   s   && r   rè   ÚDistributionWithMapper.rsample,  s>   € ð �i‰i×Ñ¤§
¢
¨<Ó 8Ó9ˆØ�;‰;Ò"Ø—+‘+˜cÓ"ˆCØˆ
r   c                óN   <€ V ^8„  d   QhRS[ S[R3,          RS[P                  /# rä   ræ   )r   r0   s   "€r   r   rß   :  s'   ø€ ÷ ñ ¡5©¨c¨¥?ð ±u·|±|ñ r   c                ó¢   € V P                   P                  \        P                  ! V4      4      pV P                  e   V P	                  V4      pV# )z•Draw a sample and apply ``map_fn`` when provided.

Args:
    sample_shape: Desired sample shape.

Returns:
    Sample tensor after optional mapping.
)rÝ   ÚsamplerR   ré   rÞ   rê   s   && r   rî   ÚDistributionWithMapper.sample:  s>   € ð �i‰i×ÑœuŸzšz¨,Ó7Ó8ˆØ�;‰;Ò"Ø—+‘+˜cÓ"ˆCØˆ
r   c                ó:   <€ V ^8„  d   QhRS[ RS[P                  /# )r   Únr   )rÑ   rR   r~   )r   r0   s   "€r   r   rß   H  s   ø€ ÷ ñ ™#ð ¡%§,¡,ñ r   c                óz   € V P                   P                  V4      pV P                  e   V P                  V4      pV# )zŒDraw ``n`` samples and apply ``map_fn`` when provided.

Args:
    n: Number of samples.

Returns:
    Sample tensor after optional mapping.
)rÝ   Úsample_nrÞ   )rH   rñ   rÇ   s   && r   ró   ÚDistributionWithMapper.sample_nH  s5   € ð �i‰i× Ñ  Ó#ˆØ�;‰;Ò"Ø—+‘+˜cÓ"ˆCØˆ
r   c                ó&   <€ V ^8„  d   QhRS[ RS[/# )r   Úattrr   )re   r   )r   r0   s   "€r   r   rß   V  s   ø€ ÷ Dñ D¡ð D©ñ Dr   c                ó´   € VP                  R 4      '       d#   VP                  R 4      '       d   \        V4      h\        \        P                  V R4      V4      # )Ú__rÝ   )Ú
startswithÚendswithÚAttributeErrorÚgetattrÚobjectÚ__getattribute__)rH   rö   s   &&r   Ú__getattr__Ú"DistributionWithMapper.__getattr__V  sF   € ð �?‰?˜4× Ò  T§]¡]°4×%8Ò%8Ü  Ó&Ð&Ü”v×.Ñ.¨t°VÓ<¸dÓCÐCr   rá   r   )r9   r:   r;   r<   r=   rG   rè   rî   ró   rÿ   r>   r?   r@   s   @r   rÛ   rÛ     s>   ø‡ € ñ÷2ò ÷ð ÷ð ÷ð ÷Dö Dr   rÛ   c                   ó   € ] tR tRtRtRtR# )ÚUniformDistributioni^  zgWrapper around torch Uniform distribution which makes it work with the 'spawn' multiprocessing context.r8   N)r9   r:   r;   r<   r=   r>   r8   r   r   r  r  ^  s   † Ýqr   r  )Útypingr   r   r   r   r   r   r	   r
   r   rR   r   Útorch.distributionsr   r   Ú!kornia.augmentation.utils.helpersr   r   Ú*kornia.augmentation.utils.param_validationr   r   r   rý   r[   r*   r3   r,   ÚModuler   rÛ   r  r8   r   r   Ú<module>r     sz   ð÷$ S× RÕ Rã Ý ß 5ç Vß _áˆCƒL€á	‹€õ	Gô $ô ôV"˜"Ÿ)™)Ð/Jõ V"ôrMD˜\ô MDô`rÐ-¨wö rr   