+
    þäûj:L  ã                   óš  a € R) t$0 t ^ RIt^ RIHt ^ RIHtHtHtHtH	t	H
t
 ^ RIt^ RIHt ^ RIHt ]P                   ! 4       tRs] ^ k R R ltR R	 ltR
 R ltR R ltR R ltR R ltR R ltR*R R lltR^ ]! R4      3RRR3R R lltR^ ]! R4      33R R lltR R ltR R  ltR! R" lt R*R# R$ llt!R+R% R& llt"R,R' R( llt#R# )-é    N)Úpartial)ÚAnyÚDictÚListÚOptionalÚTupleÚUnion)Ú_constant_tensor)Úis_compilingc                óR   € V ^8„  d   QhR\         \        P                  ,          RR/# )é   ÚdeviceÚreturnN©r   Útorchr   )Úformats   "ÚfC:\Ai_Coding\apps\studio\ComfyUI\.venv\Lib\site-packages\kornia\augmentation\utils\param_validation.pyÚ__annotate__r   &   s#   € ÷ (ñ (¤¬¯©Õ 6ð (¸4ñ (ó    c                 óD   € \        4       '       d   V sR # V \        n        R # ©N)r   Ú_compiling_rebuild_deviceÚ_thread_rebuildr   )r   s   &r   Ú_set_rebuild_devicer   &   s   € Ü‡~‚~à$*Ò!à!'ŒÖr   c                óN   € V ^8„  d   QhR\         \        P                  ,          /# ©r   r   r   )r   s   "r   r   r   .   s   € ÷ 4ñ 4œX¤e§l¡lÕ3ñ 4r   c                  óP   € \        4       '       d   \        # \        \        R R4      # )r   N)r   r   Úgetattrr   © r   r   Ú_get_rebuild_devicer    .   s   € Ü‡~‚~Ü(Ð(Ü”? H¨dÓ3Ð3r   c                ó$   € V ^8„  d   QhR\         /# r   )Úbool)r   s   "r   r   r   4   s   € ÷ -ñ -œTñ -r   c                  ó   € \        4       R J# r   )r    r   r   r   Ú_in_sampler_rebuildr$   4   s   € ÜÓ ¨Ð,Ð,r   c                ó°   € V ^8„  d   QhR\         R\        \        P                  ,          R\        \        P                  ,          R\        P
                  /# ©r   Údatar   Údtyper   ©r   r   r   r   r(   ÚTensor)r   s   "r   r   r   8   sD   € ÷ >ñ >œSð >¬(´5·<±<Õ*@ð >ÌÔRW×R]ÑR]ÕI^ð >Ôch×coÑcoñ >r   c                ót   € \        4       '       d   Vf   \        P                  ! WVR7      # \        WVR7      # )zgReturn ``torch.tensor(data, device=device, dtype=dtype)``, without a host-device copy during a rebuild.©r   r(   )r$   r   Útensorr
   ©r'   r   r(   s   &&&r   Ú_numbers_to_tensorr/   8   s-   € ä× Ò  E¢MÜ�|Š|˜D°uÔ=Ð=Ü˜D°uÔ=Ð=r   c                óš   € V ^8„  d   QhR\         R\        \        P                  ,          R\        P                  R\        P
                  /# r&   r)   )r   s   "r   r   r   ?   s<   € ÷ >ñ >œ3ð >¬´·±Õ(>ð >ÄuÇ{Á{ð >ÔW\×WcÑWcñ >r   c                ó¬   € \        4       '       d!   \        V \        P                  4      '       d   \        P                  ! WVR7      # \        WVR7      # )zjReturn ``torch.as_tensor(data, device=device, dtype=dtype)``, without a host-device copy during a rebuild.r,   )r$   Ú
isinstancer   r*   Ú	as_tensorr
   r.   s   &&&r   Ú_as_range_tensorr4   ?   s9   € ä× Ò ¤J¨t´U·\±\×$BÒ$BÜ�Š˜t¸%Ô@Ð@Ü˜D°uÔ=Ð=r   c                ó4   € V ^8„  d   QhR\         R\        RR/# )r   ÚvalueÚnamer   N)r   Ústr)r   s   "r   r   r   F   s&   € ÷ Sñ S¬ð S´Cð S¸Dñ Sr   c                óŠ   € \        V 4      '       d   R# \        V \        4      '       d   V ^ 8”  g   \        RV RV  R24      hR# )aP  Assert ``value`` is a non-negative Python int, or accept it as a traced dim.

Used by random generators to validate ``height``/``width``/``depth`` derived from
``tensor.shape``. Under tracing those are 0-d tensors; we accept them rather than
aborting export, since the surrounding ops will fail loudly if the value is
actually invalid.
NÚ`z"` must be a positive integer. Got Ú.)Ú_is_traced_dimr2   ÚintÚAssertionError)r6   r7   s   &&r   Ú_check_positive_int_or_tracedr?   F   sE   € ô �e×ÒÙÜ�uœc×"Ò" u¨q¤yÜ˜q  Ð&HÈÈÈqÐQÓRÐRñ (1r   c                ó0   € V ^8„  d   QhR\         R\        /# )r   r6   r   )r   r"   )r   s   "r   r   r   T   s   € ÷ 	Wñ 	Wœ#ð 	W¤$ñ 	Wr   c                ó¢   € \        V \        P                  4      ;'       d/    V P                  4       ^8H  ;'       d    V P	                  4       ^8*  # )a²  Return True if ``value`` looks like a shape dim that came from tensor tracing.

Under torch.onnx.export's TorchScript tracer (and torch.fx), accessing
``tensor.shape[i]`` can yield a 0-d tensor instead of a Python ``int``. The
runtime validators in this module are redundant during tracing, the tensor
constructors that follow will catch any real shape problem. Treat such
"traced ints" as opaque-but-valid so the trace can continue.
)r2   r   r*   ÚnumelÚdim)r6   s   &r   r<   r<   T   s<   € ô �eœUŸ\™\Ó*×VÐV¨u¯{©{«}ÀÑ/A×VÐVÀeÇiÁiÃkÐUVÑFVÐVr   c                óJ   € V ^8„  d   QhR\         R\        \        ,          RR/# )r   Ú
batch_sizeÚsame_on_batchr   N)r=   r   r"   )r   s   "r   r   r   `   s+   € ÷ Xñ X¤Cð X¼Ä½ð XÐRVñ Xr   c                óÚ   € \        V 4      '       d   R# \        V \        4      '       d   V ^ 8¼  g   \        RV  R24      hVe(   \        V\        4      '       g   \        RV R24      hR# R# )zìCheck valid batch_size and same_on_batch params.

Skips the integer-type check when ``batch_size`` is a 0-d / 1-element tensor
(the shape of a traced tensor), since static configuration is already
validated at module construction time.
Nz.`batch_size` shall be a positive integer. Got r;   z&`same_on_batch` shall be boolean. Got )r<   r2   r=   r>   r"   )rE   rF   s   &&r   Ú_common_param_checkrH   `   sl   € ô �j×!Ò!ÙÜ�z¤3×'Ò'¨J¸!¬OÜÐMÈjÈ\ÐYZÐ[Ó\Ð\ØÒ ¬°MÄ4×)HÒ)HÜÐEÀmÀ_ÐTUÐVÓWÐWñ *IÑ r   g        ÚinfÚjointc                óÆ  € V ^8„  d   QhR\         \        P                  \        \        \        \        3,          \
        \        ,          3,          R\        R\        \        ,          R\        \        \        \        3,          ,          R\        \        ,          R\        \        P                  ,          R\        \        P                  ,          R\        P                  /# )	r   Úfactorr7   ÚcenterÚboundsÚcheckr   r(   r   )
r	   r   r*   Úfloatr   r   r8   r   r   r(   )r   s   "r   r   r   o   s­   € ÷ Eñ EÜ”%—,‘,¤¤u¬U´E¨\Õ':¼DÄ½KÐGÕHðEä
ðEô ”U�OðEô ”Uœ5¤%˜<Õ(Õ)ð	Eô
 ”C�=ðEô ”U—\‘\Õ"ðEô ”E—K‘KÕ ðEô ‡\�\ñEr   c                ót  € \        4       pVRJ pVf   Vf   \        P                  ! R4      MTpVf   \        P                  ! 4       p\	        V \        P
                  4      '       g   \        WV4      p V P                  4       ^ 8X  Ed   V'       dH   V ^ 8  d   \        RV RV  R24      h\        P                  ! V 4      '       g   \        RV RV  R24      hVe   Vf   \        RV RV R24      hV P                  ^4      \        RR	.V P                  V P                  4      ,          V,           p	V'       d*   V	^,          V^,          8”  d   \        V R
V RV	 R24      hV	P                  V^ ,          R7      P                  WVR7      p	Me\        P                  ! WVR7      p	V'       dF   \        \        P                  ! V	4      P!                  4       4      '       g   \        RV RV	 R24      hV'       d>   Ve:   VR8X  d   \#        W‘V4       V	# VR8X  d   \%        W‘V4       V	# \'        RV R24      hV	# )a  Check inputs and compute the corresponding factor bounds.

A scalar ``factor`` is a magnitude around ``center``: the range is ``[center - factor, center + factor]``
with its lower end floored at ``bounds[0]`` (a non-negative parameter reads ``contrast=1.5`` as
``[0, 2.5]``, the way torchvision does), and an upper end past ``bounds[1]`` raises the same
``ValueError`` the explicit range would. The scalar must be finite: ``inf``, ``nan``, and a finite
magnitude too large to represent in ``dtype`` (which converts to ``inf``) all raise ``ValueError``,
while a negative value reports "must be non negative" from the earlier check. A pair is taken as the
range itself, must likewise be finite, and is checked against ``bounds`` with ``check``. Finiteness
is required of the range, not of ``bounds``: the domain is legitimately unbounded (the default is
``(0, inf)``), but a draw range that is not finite produces ``NaN`` at forward time.

During a rebuild of the samplers none of the values are checked (see ``_thread_rebuild``).
NÚcpuzIf z2 is a single number, it must be non negative. Got r;   z, is a single number, it must be finite. Got z<`center` and `bounds` cannot be None for single number. Got z, g      ð?ú  out of bounds. Expected inside ú, got )Úminr,   z$ is a range, it must be finite. Got rJ   Úsingularz	methods 'z' not implemented.g      ð¿)r    r   r   Úget_default_dtyper2   r*   r/   rC   Ú
ValueErrorÚisfiniteÚrepeatr(   ÚclampÚtor3   r"   ÚallÚ_joint_range_checkÚ_singular_range_checkÚNotImplementedError)
rL   r7   rM   rN   rO   r   r(   Úrebuild_deviceÚcheckedÚfactor_bounds
   &&&&&&&   r   Ú_range_boundrd   o   s&  € ô. )Ó*€NØ Ð$€GØ‚~ð )7Ò(>”—’˜eÔ$ÀNˆØ‚}Ü×'Ò'Ó)ˆÜ�fœuŸ|™|×-Ò-Ü# F°EÓ:ˆð ‡z�zƒ|�qÕßØ˜ŒzÜ  3 t fÐ,^Ð_eÐ^fÐfgÐ!hÓiÐiÜ—>’> &×)Ò)Ü  3 t fÐ,XÐY_ÐX`Ð`aÐ!bÓcÐcØŠ>˜Vš^ÜÐ[Ð\bÐ[cÐceÐflÐemÐmnÐoÓpÐpØ—}‘} QÓ'Ô*<¸dÀC¸[È&Ï-É-ÐY_×YeÑYeÓ*fÕfÐioÕoˆ÷ �| A•¨°­Ô2Ü ˜vÐ%EÀfÀXÈVÐT`ÐSaÐabÐcÓdÐdØ#×)Ñ)¨f°Q­iÐ)Ó8×;Ñ;À6Ð;ÓW‰ä—’ vÀEÔJˆ÷ œ4¤§¢¨|Ó <× @Ñ @Ó B×CÒCÜ˜s 4 &Ð(LÈ\ÈNÐZ[Ð\Ó]Ð]ç�5Ò$Ø�GÔÜ˜|°6Ô:ð Ðð �jÔ Ü! ,°fÔ=ð Ðô &¨	°%°Ð8JÐ&KÓLÐLàÐr   c                ó²  € V ^8„  d   QhR\         \        P                  \        \        \        \        3,          \
        \        ,          3,          R\        R\        \        ,          R\        \        \        \        3,          ,          R\        \        P                  \        \        ,          \        \        \        \        3,          ,          3,          /# )r   rL   r7   rM   rN   r   )r	   r   r*   rP   r   r   r8   r   )r   s   "r   r   r   ·   s�   € ÷ ñ Ü”%—,‘,¤¤u¬U´E¨\Õ':¼DÄ½KÐGÕHðä
ðô ”U�Oðô ”Uœ5¤%˜<Õ(Õ)ð	ô
 Œ5�<‰<œ¤%�¬(´5¼Ä¸Õ3FÕ*GÐGÕHñr   c                óê   € \        V \        P                  P                  4      '       g   \	        WW#R7      RR3# \	        V P                  4       WVR7       V P                  4       ^ 8X  d   WV3# V RR3# )a^  Check a sampling range like :func:`_range_bound`, keeping a learnable one as the leaf that was passed.

Returns the range to keep on the module, then the ``center`` and ``bounds`` that ``PlainUniformGenerator`` still
has to apply when it builds its samplers. :func:`_range_bound` converts a range to the default device and dtype and
derives ``[center - factor, center + factor]`` from a number, which turns an ``nn.Parameter`` into a tensor that the
module does not own and a copy cannot take. A parameter is therefore checked on a detached copy and kept as it is:
a ``(2,)`` one is the range itself, and the range of a 0-d one is derived in ``make_samplers``, on the generator's
device and dtype, from the parameter of whichever module builds the samplers. Any other input returns the
:func:`_range_bound` result with no ``center`` or ``bounds`` left to apply.
)rM   rN   N)r2   r   ÚnnÚ	Parameterrd   ÚdetachrC   )rL   r7   rM   rN   s   &&&&r   Ú_learnable_range_boundrj   ·   sg   € ô  �fœeŸh™h×0Ñ0×1Ò1Ü˜F°ÔGÈÈtÐSÐSÜ�—‘“ $¸fÕEØ‡z�zƒ|�qÔØ˜vÐ%Ð%Ø�4˜ÐÐr   c                ó~   € V ^8„  d   QhR\         P                  P                  R\        R\        \
        ,          RR/# )r   Úmoduler7   rM   r   N)r   rg   ÚModuler8   r   rP   )r   s   "r   r   r   Ï   s<   € ÷ hñ h¬%¯(©(¯/©/ð hÄð hÌhÔW\Íoð hÐbfñ hr   c                óÎ   € \        V P                  P                  V4      \        P                  P
                  4      '       d"   V P                  \        \        W4      4       R# R# )a  Let ``module`` strict-load a checkpoint saved when it did not own its learnable range ``name`` yet.

Use it after :func:`_learnable_range_bound`, with the ``center`` it returned. The colour augmentations used to keep
an ``nn.Parameter`` range only when it was a ``(2,)`` CPU tensor in the default dtype. Their other checkpoints hold
no ``name`` entry, only the generator's ``<name>_factor``: the range as a ``(2,)`` tensor, and for a 0-d parameter
the range derived from it, ``[max(center - f, low), center + f]``. A load pre-hook fills ``name`` from that entry,
and for a 0-d parameter recovers ``f`` as the upper end minus ``center``, because the upper end is never clamped.
Nothing is registered for a range that is not a parameter, whose checkpoint layout did not change.
N)	r2   Ú_parametersÚgetr   rg   rh   Ú"_register_load_state_dict_pre_hookr   Ú_fill_unowned_learnable_range)rl   r7   rM   s   &&&r   Ú_load_unowned_learnable_rangers   Ï   sI   € ô �&×$Ñ$×(Ñ(¨Ó.´·±×0BÑ0B×CÒCØ×1Ñ1´'Ô:WÐY]Ó2fÖgñ Dr   c                ó�   € V ^8„  d   QhR\         R\        \        ,          R\        \         \        3,          R\         R\        RR/# )r   r7   rM   Ú
state_dictÚprefixÚargsr   N)r8   r   rP   r   r   )r   s   "r   r   r   Ý   sH   € ÷ ñ Ü
ðÜ¤•ðÜ48¼¼c¸µNðÜLOðÜX[ðà	ñr   c                 ó  € W0,           V R V  R2reWR9   g   Wb9  d   R# W&,          pVeT   \        V\        P                  4      '       d4   VP                  R8X  d#   VP	                  4       ^,          V,
          pWrV&   WrV&   R# )z_param_generator.Ú_factorN)r   )r2   r   r*   Úshaperi   )r7   rM   ru   rv   rw   ÚkeyÚ
factor_keyÚsaveds   &&&&*   r   rr   rr   Ý   sw   € ð •m¨ xÐ/@ÀÀÀgÐ%NˆØ
Ô˜JÔ8ÙØÕ"€EàÒœj¨´·±×=Ò=À%Ç+Á+ÐQUÔBUØ—‘“˜qÕ! FÕ*ˆØ!&�:ÑØˆsƒOr   c          
      ó&  € V ^8„  d   QhR\         \        P                  \        \        \        \        3,          \        \        \        \        \        3,          3,          R\        P
                  R\        P                  R\        P                  /# )r   Úshearr   r(   r   )r	   r   r*   rP   r   r   r(   )r   s   "r   r   r   ë   sf   € ÷ +ñ +Ü”—‘œu¤e¬E´5¨LÕ&9¼5ÄÌÌuÔV[ÐA[Õ;\Ð\Õ]ð+ä�L‰Lð+ô �;‰;ð+ô ‡\�\ñ	+r   c           	     ó&  € \        WV4      p V P                  \        P                  ! ^^.4      8X  dF   \	        V ^ ,          R^ RWR7      p\	        V ^,          R^ RWR7      p\        P
                  ! W4.4      # \	        V P                  4       ^ 8X  d   T MV R,          R^ RWR7      pV P                  4       ^ 8X  g   \        V 4      ^8X  d   \        P                  ! ^WR7      pM\	        V R,          R^ RWR7      p\        P
                  ! W4.4      # )zåReturn the x- and y-axis shear ranges as a (2, 2) tensor on ``device`` in ``dtype``.

A float or an (a, b) pair only shears along the x axis, (a, b, c, d) also gives a y-axis
range, and a (2, 2) tensor already holds both ranges.
zshear-xr,   zshear-y:Nr   N:r   NN)i˜þÿÿih  )	r4   rz   r   ÚSizerd   ÚstackrC   ÚlenÚzeros)r   r   r(   Úshear_xÚshear_ys   &&&  r   Ú_shear_boundr‡   ë   sâ   € ô ˜U¨EÓ2€EØ‡{�{”e—j’j ! Q Ó(Ô(Ü˜u Q�x¨°A°{È6Ô_ˆÜ˜u Q�x¨°A°{È6Ô_ˆÜ�{Š{˜GÐ-Ó.Ð.äØ—‘“ Ô!‰ u¨R¥y°)¸QÀÐTZô€Gð ‡y�yƒ{�aÔœ3˜u›:¨œ?Ü—+’+˜a¨Ô<‰ä˜u R�y¨)°Q¸ÈFÔ`ˆÜ�;Š;˜Ð)Ó*Ð*r   c          	      óŒ   € V ^8„  d   QhR\         P                  R\        R\        \        \
        \
        3,          ,          RR/# )r   Úranged_factorr7   rN   r   N)r   r*   r8   r   r   rP   )r   s   "r   r   r     s?   € ÷ 
ñ 
¤e§l¡lð 
¼#ð 
ÄxÔPUÔV[Ô]bÐVbÕPcÕGdð 
Ðptñ 
r   c                óÜ  € \        4       '       d   R# Vf   \        R4      \        R4      3pV P                  4       ^8X  d™   \        V 4      ^8X  d‰   V^ ,          V ^ ,          8:  d   V^,          V ^,          8¼  g   \	        V RV RV  R24      hV^ ,          V ^ ,          u;8:  d   V ^,          u;8:  d   V^,          8:  g   M \	        V RV RV  24      hR# \        V R	V R
V  R24      h)zHCheck if bounds[0] <= ranged_factor[0] <= ranged_factor[1] <= bounds[1].Nú-infrI   rS   rT   r;   z[0] should be smaller than z[1] got z= should be a torch.Tensor with length 2 whose values between z. Got )r$   rP   rC   rƒ   rX   Ú	TypeError)r‰   r7   rN   s   &&&r   r^   r^     sø   € ä×ÒÙØ‚~Ü˜“-¤ u£Ð.ˆØ×ÑÓ˜aÔ¤C¨Ó$6¸!Ô$;Ø�a�y˜M¨!Õ,Ô,°F¸1µIÀÈqÕAQÔ4QÜ ˜vÐ%EÀfÀXÈVÐTaÐSbÐbcÐdÓeÐeà�a�y˜M¨!Õ,ÖM°¸aÕ0@ÖMÀFÈ1ÅIÖMÜ ˜vÐ%@ÀÀÀhÈ}ÈoÐ^Ó_Ð_ñ Nô ØˆfÐQÐRXÐQYÐY_Ð`mÐ_nÐnoÐpó
ð 	
r   c                ó¤   € V ^8„  d   QhR\         P                  R\        R\        \        \
        \
        3,          ,          R\        R\        RR/# )r   r‰   r7   rN   Ú	skip_noneÚmoder   N)r   r*   r8   r   r   rP   r"   )r   s   "r   r   r     sV   € ÷ 
ñ 
Ü—<‘<ð
ä
ð
ô ”Uœ5¤%˜<Õ(Õ)ð
ô ð	
ô
 ð
ð 
ñ
r   c           	     ó¬  € \        4       '       d   R# VR8X  d   ^pMVR8X  d   ^pM\        RV 24      hV'       d   V f   R# Vf   \        R4      \        R4      3pV P                  4       ^8X  dO   \	        V 4      V8X  d?   V  F6  pV^ ,          Tu;8:  d   V^,          8:  d   K"  M \        V RV RV  R	24      h	  R# \        V R
V RV RV  24      h)zcCheck if bounds[0] <= ranged_factor[0] <= bounds[1] and bounds[0] <= ranged_factor[1] <= bounds[1].NÚ2dÚ3dz%'mode' shall be either 2d or 3d. Got r‹   rI   rS   rT   r;   z1 should be a float number or a tuple with length z whose values between z.Got )r$   rX   rP   rC   rƒ   rŒ   )r‰   r7   rN   rŽ   r�   Údim_sizeÚfs   &&&&&  r   r_   r_     só   € ô ×ÒÙØˆt„|Ø‰Ø	�ŒØ‰äÐ@ÀÀÐGÓHÐHç�]Ò*ÙØ‚~Ü˜“-¤ u£Ð.ˆØ×ÑÓ˜aÔ¤C¨Ó$6¸(Ô$BÛˆAØ˜!•9 Ö. V¨A¥Y×.Ð.Ü  D 6Ð)IÈ&ÈÐQWÐXeÐWfÐfgÐ!hÓiÐió ô ØˆfÐEÀhÀZÐOeÐflÐemð nØ �/ð#ó
ð 	
r   c                óf  € V ^8„  d   QhR\         \        P                  \        \        \
        R3,          3,          R\        R\        \        P                  ,          R\        \        P                  ,          R\        R\        \        \        \        3,          ,          R\        P                  /# )	r   Úinput_range.Útarget_sizer   r(   r7   rN   r   )r	   r   r*   rP   r   r   r=   r   r   r(   r8   )r   s   "r   r   r   7  s“   € ÷ Xñ XÜ”u—|‘|¤U¬E´#°s°(­OÐ;Õ<ðXäðXô ”U—\‘\Õ"ðXô ”E—K‘KÕ ð	Xô
 ðXô ”Uœ5¤%˜<Õ(Õ)ðXô ‡\�\ñXr   c           	     óP  € \         P                  ! V^.4      p\        V \         P                  4      '       EdÄ   \	        V P
                  4      ^ 8X  g+   \	        V P
                  4      ^8X  d…   \	        V 4      ^8X  du   \        4       '       g   V ^ 8  d   \        RV  24      hV P                  ^4      P                  W#R7      \        R^.W#4      ,          pVP                  V^ ,          ^4      pEM\	        V P
                  4      ^8X  d<   \	        V 4      ^8X  d,   V P                  V^ ,          ^4      P                  W#R7      pEMÁ\	        V P
                  4      ^8X  d^   \	        V 4      V^ ,          8X  dG   V P                  ^4      P                  ^^4      P                  W#R7      \        R^.W#4      ,          pEMJV P
                  V8X  d   V P                  W#R7      pEM%\        R\        V4       RV  24      h\        V \        \        34      '       d>   V ^ 8  d   \        RV  24      h\        V ) V .W#4      P                  V^ ,          ^4      pEM²\        V \        \        34      '       d|   \	        V 4      ^8X  dl   \        V ^ ,          \        \        34      '       dI   \        V ^,          \        \        34      '       d&   \        WV4      P                  V^ ,          ^4      pEM\        V \        \        34      '       dq   \	        V 4      V^ ,          8X  dZ   \         ;QJ d    R V  4       F  '       d   K   RM	  RM! R V  4       4      '       d    \        V  Uu. uF  qˆ) V3NK	  	  upW#4      pM�\        V \        \        34      '       d_   \	        V 4      V^ ,          8X  dH   \         ;QJ d    R V  4       F  '       d   K   RM	  RM! R V  4       4      '       d   \        WV4      pM\#        R	V R
V RV  R24      hVe   V F  p	\%        W”V4       K  	  V# u upi )aQ  Given target_size, it will generate the corresponding (target_size, 2) range torch.Tensor.

This is for element-wise params.

Every read range is checked against ``bounds``, when given, the way :func:`_range_bound` checks
the 2D scalar and explicit forms, so a 3D angle past one turn raises at construction instead of
being sampled (#4617). ``name`` only names the parameter in that error.

Example:
>>> degree = torch.tensor([0.2, 0.3])
>>> _tuple_range_reader(degree, 3)  # read degree for yaw, pitch and roll.
tensor([[0.2000, 0.3000],
        [0.2000, 0.3000],
        [0.2000, 0.3000]])

zCIf input_range is only one number it must be a positive number. Gotr,   zDegrees must be a zA torch.tensor for the degree range for independent operation.Got c              3   óN   "  € T F  p\        V\        \        34      x € K  	  R # 5ir   )r2   rP   r=   ©Ú.0Úxs   & r   Ú	<genexpr>Ú&_tuple_range_reader.<locals>.<genexpr>z  s   é € ÐA±[°”
˜1œu¤c˜l×+Ð+³[ùó   ‚#%FTc              3   óN   "  € T F  p\        V\        \        34      x € K  	  R # 5ir   )r2   ÚtupleÚlistrš   s   & r   r�   rž   �  s   é € ÐB±k°”
˜1œu¤d˜m×,Ð,³kùrŸ   zcIf not pass a torch.tensor, it must be float, (float, float) for isotropic operation or a tuple of z floats or z/ (float, float) for independent operation. Got r;   éÿÿÿÿ)r   r�   r2   r*   rƒ   rz   r$   rX   rZ   r\   r/   Ú	unsqueezer¢   rP   r=   r¡   r]   rŒ   r^   )
r–   r—   r   r(   r7   rN   Útarget_shapeÚinput_range_tmpÚsÚrows
   &&&&&&    r   Ú_tuple_range_readerr©   7  s�  € ô0 —:’:˜{¨AÐ.Ó/€Lä�+œuŸ|™|×,Ó,Ü�×!Ñ!Ó" aÔ'¬S°×1BÑ1BÓ-CÀqÔ-HÌSÐQ\ÓM]ÐabÔMbÜ&×(Ò(¨[¸1¬_Ü Ð#fÐgrÐfsÐ!tÓuÐuØ)×0Ñ0°Ó3×6Ñ6¸fÐ6ÓRÔUgØ�Q�˜óVõ ˆOð .×4Ñ4°\À!µ_ÀaÓHŠOä�×"Ñ"Ó# qÔ(¬S°Ó-=ÀÔ-BØ)×0Ñ0°¸aµÀ!ÓD×GÑGÈvÐGÓcŠOä�×"Ñ"Ó# qÔ(¬S°Ó-=ÀÈaÅÔ-PØ)×3Ñ3°AÓ6×=Ñ=¸aÀÓC×FÑFÈfÐFÓbÔewØ�Q�˜ófõ ŠOð ×Ñ ,Ô.Ø)Ÿn™n°F˜nÓHŠOô Ø$¤T¨,Ó%7Ð$8ð 9Ø"�mð%óð ô 
�K¤%¬ ×	.Ò	.Ø˜Œ?ÜÐbÐcnÐboÐpÓqÐqÜ,¨{¨l¸KÐ-HÈ&ÓX×_Ñ_Ð`lÐmnÕ`oÐqrÓsŠô 	�;¤¬ ×.Ò.Ü�Ó Ô!Ü�{ 1•~¬¬s |×4Ò4Ü�{ 1•~¬¬s |×4Ò4ä,¨[À%ÓH×OÑOÐP\Ð]^ÕP_ÐabÓcŠô 	�;¤¬ ×.Ò.Ü�Ó ¨Q¥Ô/ß‹CÑA±[ÓA�C�CŠCÑA±[ÓA×AÒAä,¹{Ó-K¹{¸!¨r°1«g¹{Ñ-KÈVÓ[‰ô 	�;¤¬ ×.Ò.Ü�Ó ¨Q¥Ô/ß‹CÑB±kÓB�C�CŠCÑB±kÓB×BÒBä,¨[À%ÓH‰ô ØqØˆm˜; { mÐ3bÐcnÐboÐopðró
ð 	
ð
 ÒÛ"ˆCÜ˜s¨&Ö1ñ #ð Ðùò' .Ls   Í#P#c                ój   € V ^8„  d   Qh/ ^ \         9   d   \        \        P                  ,          ;R&   # )r   r   )Ú__conditional_annotations__r   r   r   )r   s   "r   r   r      s#   € × Ñ ÷" 9Ò 8œ8¤E§L¡LÕ1Ñ 8ò# r   r   )NFr‘   )NNr–   N)%r«   Ú	threadingÚ	functoolsr   Útypingr   r   r   r   r   r	   r   Úkornia.core.opsr
   Úkornia.core.utilsr   Úlocalr   r   r   r    r$   r/   r4   r?   r<   rH   rP   rd   rj   rs   rr   r‡   r^   r_   r©   r   )r«   s   @r   Ú<module>r²      sÀ   øð÷$ Ð Ý ß :× :ã å ,Ý *ð —/’/Ó#€Ø48Ð Ó 8õ(õ4õ-õ>õ>õSõ	W÷Xð$ "Ø-.±°e³Ð,=Ø"Ø%)Ø#'÷EðV "Ø-.±°e³Ð,=÷	õ0hõõ+÷4
÷$
÷@Xñ Xr   