+
    åûjö5  ã                  óæ  a € 0 t $ ^ RIHt ^ RIt^ RIHt ^ RIHtHt ^ RI	t	^ RI
Hu Ht ^ RI	Ht ^ RIHt ^ RIHtHt ^ RIHt R	.t]! R
R4      R.tR]R&   R R lt ! R R]P2                  4      t ! R R]P6                  4      t ! R R]P6                  4      t ! R R]P2                  4      t ! R R]P2                  4      t ! R R]P6                  4      t  ! R R	]]P2                  4      t!R# )é    )ÚannotationsN)ÚOrderedDict)ÚClassVarÚOptional)Únn)ÚKORNIA_CHECK)Úhf_urlÚload_state_dict_from_url)ÚONNXExportMixinÚDexiNedÚdexinedzDexiNed_BIPED_10.pthz>https://cmp.felk.cvut.cz/~mishkdmy/models/DexiNed_BIPED_10.pthústr | list[str]Ú_urlc               ó    € V ^8„  d   QhRRRR/# )é   Úmz	nn.ModuleÚreturnÚNone© )Úformats   "ÚQC:\Ai_Coding\apps\studio\ComfyUI\.venv\Lib\site-packages\kornia\models\dexined.pyÚ__annotate__r   *   s   € ÷ )ñ )�9ð ) ñ )ó    c                ó>  € \        V \        P                  4      '       dí   \        P                  P                  P                  V P                  RR7       V P                  P                  P                  ^,          \        P                  ! ^.4      8X  d6   \        P                  P                  P                  V P                  RR7       V P                  e4   \        P                  P                  P                  V P                  4       \        V \        P                  4      '       dñ   \        P                  P                  P                  V P                  RR7       V P                  P                  P                  ^,          \        P                  ! ^.4      8X  d6   \        P                  P                  P                  V P                  RR7       V P                  e6   \        P                  P                  P                  V P                  4       R# R# R# )zInitialize weights.g      ð?)Úgaing        )ÚmeanNgš™™™™™¹?)Ústd)Ú
isinstancer   ÚConv2dÚtorchÚinitÚxavier_normal_ÚweightÚdataÚshapeÚSizeÚnormal_ÚbiasÚzeros_ÚConvTranspose2d)r   s   &r   Úweight_initr+   *   sB  € ä�!”R—Y‘Y×Òä�‰�‰×$Ñ$ Q§X¡X°CÐ$Ô8à�8‰8�=‰=×Ñ˜qÕ!¤U§Z¢Z°°£_Ô4Ü�H‰H�M‰M×!Ñ! !§(¡(°Ð!Ô5à�6‰6ÒÜ�H‰H�M‰M× Ñ  §¡Ô(ô �!”R×'Ñ'×(Ò(ä�‰�‰×$Ñ$ Q§X¡X°CÐ$Ô8ð �8‰8�=‰=×Ñ˜qÕ!¤U§Z¢Z°°£_Ô4Ü�H‰H�M‰M×!Ñ! !§(¡(°Ð!Ô4Ø�6‰6ÒÜ�H‰H�M‰M× Ñ  §¡Ö(ñ ñ )r   c                  ó:   a € ] tR t^BtR V 3R lltR R ltRtV ;t# )ÚCoFusionc               ó$   € V ^8„  d   QhRRRRRR/# )r   Úin_chÚintÚout_chr   r   r   )r   s   "r   r   ÚCoFusion.__annotate__C   s!   € ÷ /ñ /˜cð /¨3ð /°4ñ /r   c                	óˆ  <€ \         SV `  4        \        P                  ! V^@^^^R7      V n        \        P                  ! ^@^@^^^R7      V n        \        P                  ! ^@V^^^R7      V n        \        P                  ! 4       V n        \        P                  ! ^^@4      V n
        \        P                  ! ^^@4      V n        R# )é@   ©Úkernel_sizeÚstrideÚpaddingN)ÚsuperÚ__init__r   r   Úconv1Úconv2Úconv3ÚReLUÚreluÚ	GroupNormÚnorm_layer1Únorm_layer2)Úselfr/   r1   Ú	__class__s   &&&€r   r:   ÚCoFusion.__init__C   s‰   ø€ Ü‰ÑÔÜ—Y’Y˜u b°aÀÈ1ÔMˆŒ
Ü—Y’Y˜r 2°1¸QÈÔJˆŒ
Ü—Y’Y˜r 6°qÀÈAÔNˆŒ
Ü—G’G“IˆŒ	ÜŸ<š<¨¨2Ó.ˆÔÜŸ<š<¨¨2Ó.ˆÖr   c               ó    € V ^8„  d   QhRRRR/# ©r   Úxútorch.Tensorr   r   )r   s   "r   r   r2   L   s   € ÷ 1ñ 1˜ð 1¨,ñ 1r   c                	óH  € V P                  V P                  V P                  V4      4      4      pV P                  V P                  V P	                  V4      4      4      p\
        P                  ! V P                  V4      ^R7      pW,          P                  ^4      R,          # )é   ©Údim):NNNN.)	r?   rA   r;   rB   r<   ÚFÚsoftmaxr=   Úsum)rC   rH   Úattns   && r   ÚforwardÚCoFusion.forwardL   sv   € à�y‰y˜×)Ñ)¨$¯*©*°Q«-Ó8Ó9ˆØ�y‰y˜×)Ñ)¨$¯*©*°TÓ*:Ó;Ó<ˆÜ�yŠy˜Ÿ™ DÓ)¨qÔ1ˆð •—‘˜qÓ! <Õ0Ð0r   )r;   r<   r=   rA   rB   r?   ©Ú__name__Ú
__module__Ú__qualname__Ú__firstlineno__r:   rR   Ú__static_attributes__Ú__classcell__©rD   s   @r   r-   r-   B   s   ø† ÷/ð /÷1ó 1r   r-   c                  ó:   a € ] tR t^VtR V 3R lltR R ltRtV ;t# )Ú_DenseLayerc               ó$   € V ^8„  d   QhRRRRRR/# )r   Úinput_featuresr0   Úout_featuresr   r   r   )r   s   "r   r   Ú_DenseLayer.__annotate__W   s!   € ÷ 
ñ 
 sð 
¸#ð 
À$ñ 
r   c                	ó^  <€ \         SV `  \        R \        P                  ! 4       3R\        P
                  ! W^^^RR7      3R\        P                  ! V4      3R\        P                  ! RR7      3R\        P
                  ! W"^^RR7      3R	\        P                  ! V4      3.4      4       R
# )Úrelu1r;   T)r6   r7   r8   r(   Únorm1Úrelu2©Úinplacer<   )r6   r7   r(   Únorm2N)r9   r:   r   r   r>   r   ÚBatchNorm2d)rC   r_   r`   rD   s   &&&€r   r:   Ú_DenseLayer.__init__W   s—   ø€ Ü‰ÑÜð œbŸgšg›iÐ(ØœbŸiši¨ÐRSÐ\]ÐghÐosÔtÐuØœbŸnšn¨\Ó:Ð;ØœbŸgšg¨dÔ3Ð4ØœbŸiši¨ÐPQÐZ[ÐbfÔgÐhØœbŸnšn¨\Ó:Ð;ðó
ö	
r   c               ó    € V ^8„  d   QhRRRR/# ©r   rH   úlist[torch.Tensor]r   r   )r   s   "r   r   ra   f   s   € ÷ %ñ %Ð+ð %Ð0Bñ %r   c                	ól   € V^ ,          V^,          r2TpV  F  pV! V4      pK  	  RWC,           ,          V.# )r   g      à?r   )rC   rH   Úx1Úx2Úx3Úmods   &&    r   rR   Ú_DenseLayer.forwardf   s:   € Ø�1•�q˜•tˆBØˆÛˆCÙ�R“ŠBñ à�r•w• Ð$Ð$r   r   rT   r[   s   @r   r]   r]   V   s   ø† ÷
ð 
÷%ó %r   r]   c                  ó:   a € ] tR t^ntR V 3R lltR R ltRtV ;t# )Ú_DenseBlockc               ó(   € V ^8„  d   QhRRRRRRRR/# )r   Ú
num_layersr0   r_   r`   r   r   r   )r   s   "r   r   Ú_DenseBlock.__annotate__o   s)   € ÷ *ñ * 3ð *¸ð *È3ð *ÐSWñ *r   c                	óš   <€ \         SV `  4        \        V4       F,  p\        W#4      pV P	                  R V^,            2V4       TpK.  	  R# )Ú
denselayerN)r9   r:   Úranger]   Ú
add_module)rC   rw   r_   r`   ÚiÚlayerrD   s   &&&&  €r   r:   Ú_DenseBlock.__init__o   sE   ø€ Ü‰ÑÔÜ�zÖ"ˆAÜ Ó=ˆEØ�O‰O˜j¨!¨a­%¨	Ð2°EÔ:Ø)ŠNó #r   c               ó    € V ^8„  d   QhRRRR/# rl   r   )r   s   "r   r   rx   v   s   € ÷ ñ Ð+ð Ð0Bñ r   c                	ó,   € TpV  F  pV! V4      pK  	  V# ©Nr   )rC   rH   Úx_outrr   s   &&  r   rR   Ú_DenseBlock.forwardv   s   € ØˆÛˆCÙ˜“JŠEñ àˆr   r   rT   r[   s   @r   ru   ru   n   s   ø† ÷*ð *÷ó r   ru   c                  óR   a € ] tR t^}tR V 3R lltR R ltR R ltR R ltR	tV ;t	# )
ÚUpConvBlockc               ó$   € V ^8„  d   QhRRRRRR/# )r   Úin_featuresr0   Úup_scaler   r   r   )r   s   "r   r   ÚUpConvBlock.__annotate__~   s!   € ÷ /ñ / Cð /°3ð /¸4ñ /r   c                	ó°   <€ \         SV `  4        ^V n        ^V n        V P	                  W4      p\        VRJR4       \        P                  ! V!  V n        R# )r   Nzlayers cannot be none)	r9   r:   Ú	up_factorÚconstant_featuresÚmake_deconv_layersr   r   Ú
SequentialÚfeatures)rC   rˆ   r‰   ÚlayersrD   s   &&& €r   r:   ÚUpConvBlock.__init__~   sM   ø€ Ü‰ÑÔØˆŒØ!#ˆÔà×(Ñ(¨Ó?ˆÜ�V 4Ð'Ð)@ÔAÜŸš vÑ.ˆŽr   c               ó$   € V ^8„  d   QhRRRRRR/# )r   rˆ   r0   r‰   r   znn.ModuleListr   )r   s   "r   r   rŠ   ‡   s!   € ÷ ñ ¨cð ¸Sð À]ñ r   c                	ó�  € \         P                  ! . 4      p. ROp\        V4       Fœ  p^V,          pWB,          pV P                  WR4      pVP	                  \         P
                  ! W^4      4       VP	                  \         P                  ! RR7      4       VP	                  \         P                  ! WˆV^VR7      4       TpKž  	  V# )r   Trf   )r7   r8   )r   r   rK   é   é   )r   Ú
ModuleListr{   Úcompute_out_featuresÚappendr   r>   r*   )	rC   rˆ   r‰   r‘   Úall_padsr}   r6   Úpadr`   s	   &&&      r   rŽ   ÚUpConvBlock.make_deconv_layers‡   s›   € Ü—’˜rÓ"ˆÚ"ˆÜ�x–ˆAØ˜X�+ˆKØÕ$ˆCØ×4Ñ4°QÓAˆLØ�M‰Mœ"Ÿ)š) K¸qÓAÔBØ�M‰Mœ"Ÿ'š'¨$Ô/Ô0Ø�M‰Mœ"×,Ò,¨\ÈÐ]^ÐhkÔlÔmØ&ŠKñ !ð ˆr   c               ó$   € V ^8„  d   QhRRRRRR/# )r   Úidxr0   r‰   r   r   )r   s   "r   r   rŠ   ”   s&   € ÷ Dñ D¨ð D°sð D¸sñ Dr   c                	ó8   € W^,
          8X  d   ^# V P                   # )rK   )r�   )rC   rž   r‰   s   &&&r   r˜   Ú UpConvBlock.compute_out_features”   s   € Ø a�<Ô'ˆqÐC¨T×-CÑ-CÐCr   c               ó$   € V ^8„  d   QhRRRRRR/# )r   rH   rI   Ú	out_shapez	list[int]r   r   )r   s   "r   r   rŠ   —   s!   € ÷ >ñ >˜ð >°)ð >Àñ >r   c                	óT   € V P                  V4      p\        P                  ! W2R R7      # )Úbilinear)Úmode)r�   rN   Úinterpolate)rC   rH   r¢   Úouts   &&& r   rR   ÚUpConvBlock.forward—   s!   € Ø�m‰m˜AÓˆÜ�}Š}˜S°*Ô=Ð=r   )r�   r�   rŒ   )
rU   rV   rW   rX   r:   rŽ   r˜   rR   rY   rZ   r[   s   @r   r†   r†   }   s"   ø† ÷/ð /õõD÷>ó >r   r†   c                  ó>   a € ] tR t^œtRR V 3R llltR R ltRtV ;t# )ÚSingleConvBlockc          
     ó,   € V ^8„  d   QhRRRRRRRRRR/# )	r   rˆ   r0   r`   r7   Úuse_bsÚboolr   r   r   )r   s   "r   r   ÚSingleConvBlock.__annotate__�   s1   € ÷ /ñ / Cð /°sð /ÀCð /ÐQUð /Ðaeñ /r   c                	ó¤   <€ \         SV `  4        W@n        \        P                  ! W^VRR7      V n        \        P                  ! V4      V n        R# )rK   T)r7   r(   N)r9   r:   Úuse_bnr   r   Úconvri   Úbn)rC   rˆ   r`   r7   r¬   rD   s   &&&&&€r   r:   ÚSingleConvBlock.__init__�   s:   ø€ Ü‰ÑÔØŒÜ—I’I˜k¸À6ÐPTÔUˆŒ	Ü—.’. Ó.ˆŽr   c               ó    € V ^8„  d   QhRRRR/# rG   r   )r   s   "r   r   r®   £   s   € ÷ ñ ˜ð ¨,ñ r   c                	ón   € V P                  V4      pV P                  '       d   V P                  V4      pV# r‚   )r±   r°   r²   )rC   rH   s   &&r   rR   ÚSingleConvBlock.forward£   s*   € Ø�I‰I�a‹LˆØ�;�;ˆ;Ø—‘˜“
ˆAØˆr   )r²   r±   r°   )TrT   r[   s   @r   rª   rª   œ   s   ø† ÷/ò /÷ó r   rª   c                  ó2   a € ] tR t^ªtRR V 3R llltRtV ;t# )ÚDoubleConvBlockc               ó0   € V ^8„  d   QhRRRRRRRRRRR	R
/# )r   rˆ   r0   Úmid_featuresr`   zOptional[int]r7   Úuse_actr­   r   r   r   )r   s   "r   r   ÚDoubleConvBlock.__annotate__«   sF   € ÷ <ñ <àð<ð ð<ð $ð	<ð
 ð<ð ð<ð 
ñ<r   c                	ó  <€ \         SV `  4        Vf   TpV P                  R\        P                  ! W^^VR7      4       V P                  R\        P
                  ! V4      4       V P                  R\        P                  ! RR7      4       V P                  R\        P                  ! W#^^R7      4       V P                  R	\        P
                  ! V4      4       V'       d*   V P                  R
\        P                  ! RR7      4       R # R # )Nr;   )r8   r7   Úbn1rc   Trf   r<   )r8   Úbn2re   )r9   r:   r|   r   r   ri   r>   )rC   rˆ   rº   r`   r7   r»   rD   s   &&&&&&€r   r:   ÚDoubleConvBlock.__init__«   sº   ø€ ô 	‰ÑÔØÒØ'ˆLØ�‰˜¤§¢¨;ÀaÐQRÐ[aÔ!bÔcØ�‰˜œrŸ~š~¨lÓ;Ô<Ø�‰˜¤§¢°Ô!6Ô7Ø�‰˜¤§¢¨<ÀqÐRSÔ!TÔUØ�‰˜œrŸ~š~¨lÓ;Ô<ßØ�O‰O˜G¤R§W¢W°TÔ%:Ö;ñ r   r   )NrK   T)rU   rV   rW   rX   r:   rY   rZ   r[   s   @r   r¸   r¸   ª   s   ø† ÷<÷ <ð <r   r¸   c                  ó|   a € ] tR t^¿t$ Rt. ROtR]R&   . ROtR]R&   R V 3R lltR R lt	R	 R
 lt
R R ltRtV ;t# )r   aÐ  Definition of the DXtrem network from :cite:`xsoria2020dexined`.

The input is a BGR image with pixel values in :math:`[0, 255]` minus the BGR mean ``[103.939, 116.779, 123.68]``:
the pretrained weights were trained on images read with ``cv2.imread`` and shifted by that mean
(xavysp/DexiNed@d944b70). The detector that :py:class:`~kornia.contrib.edge_detection.EdgeDetectorBuilder` builds
converts an RGB image in :math:`[0, 1]` to this input and applies a sigmoid to the output.

Return:
    The fused edge logits, one tensor with shape ``(B,1,H,W)``.

Example:
    >>> img = torch.rand(1, 3, 320, 320)
    >>> net = DexiNed(pretrained=False)
    >>> out = net(img)
    >>> out.shape
    torch.Size([1, 1, 320, 320])

zClassVar[list[int]]ÚONNX_DEFAULT_INPUTSHAPEÚONNX_DEFAULT_OUTPUTSHAPEc               ó    € V ^8„  d   QhRRRR/# )r   Ú
pretrainedr­   r   r   r   )r   s   "r   r   ÚDexiNed.__annotate__Ö   s   € ÷ %$ñ %$ 4ð %$¨Dñ %$r   c                	ó  <€ \         SV `  4        \        ^^ ^@^R7      V n        \        ^@^€RR7      V n        \        ^^€R4      V n        \        ^RR4      V n        \        ^RR4      V n        \        ^RR4      V n	        \        P                  ! ^^^R7      V n        \        ^@^€^4      V n        \        ^€R^4      V n        \        RR^4      V n        \        RR^4      V n        \        RR^4      V n        \        ^€R^4      V n        \        ^€R^4      V n        \        RR^4      V n        \        RR^4      V n        \        RR^4      V n        \1        ^@^4      V n        \1        ^€^4      V n        \1        R^4      V n        \1        R^4      V n        \1        R^4      V n        \1        R^4      V n        \        ^^^RR7      V n        V'       d   V PA                  \B        4       R# V PE                  \F        4       R# )	r•   )r7   F)r»   é   i   r5   )r7   r¬   N)$r9   r:   r¸   Úblock_1Úblock_2ru   Údblock_3Údblock_4Údblock_5Údblock_6r   Ú	MaxPool2dÚmaxpoolrª   Úside_1Úside_2Úside_3Úside_4Úside_5Úpre_dense_2Úpre_dense_3Úpre_dense_4Úpre_dense_5Úpre_dense_6r†   Ú
up_block_1Ú
up_block_2Ú
up_block_3Ú
up_block_4Ú
up_block_5Ú
up_block_6Ú	block_catÚload_from_filer   Úapplyr+   )rC   rÅ   rD   s   &&€r   r:   ÚDexiNed.__init__Ö   s´  ø€ Ü‰ÑÔÜ& q¨"¨b¸Ô;ˆŒÜ& r¨3¸Ô>ˆŒÜ# A s¨CÓ0ˆŒÜ# A s¨CÓ0ˆŒÜ# A s¨CÓ0ˆŒÜ# A s¨CÓ0ˆŒÜ—|’|°¸!ÀQÔGˆŒô & b¨#¨qÓ1ˆŒÜ% c¨3°Ó2ˆŒÜ% c¨3°Ó2ˆŒÜ% c¨3°Ó2ˆŒÜ% c¨3°Ó2ˆŒô +¨3°°QÓ7ˆÔÜ*¨3°°QÓ7ˆÔÜ*¨3°°QÓ7ˆÔÜ*¨3°°QÓ7ˆÔÜ*¨3°°QÓ7ˆÔô & b¨!Ó,ˆŒÜ% c¨1Ó-ˆŒÜ% c¨1Ó-ˆŒÜ% c¨1Ó-ˆŒÜ% c¨1Ó-ˆŒÜ% c¨1Ó-ˆŒÜ(¨¨A°aÀÔFˆŒ÷ Ø×Ñ¤Ö%à�J‰J”{Ö#r   c               ó    € V ^8„  d   QhRRRR/# )r   Ú	path_filer   r   r   r   )r   s   "r   r   rÆ   ý   s   € ÷ ñ ¨ð ¸Dñ r   c                ó
  € \        V\        4      '       d    \        P                  P	                  V4      MRpVeq   \        P                  P                  V4      '       dL   \        VR4      ;_uu_ 4       p\        P                  ! V\        P                  ! R4      RR7      pRRR4       M!\        V\        P                  ! R4      R7      pV P                  XRR7       V P                  4        R#   + '       g   i     L5; i)aó  Load pretrained DexiNed weights and switch the module to evaluation mode.

Args:
    path_file: path of a local checkpoint file, a URL accepted by
        :func:`kornia.core.download.load_state_dict_from_url`, or a list
        of candidate URLs tried in order (HF-first fallback). A string
        naming an existing file (a leading ``~`` is expanded) is read with
        :func:`torch.load` and ``weights_only=True``; anything else goes
        through the hub cache, which is keyed by base name.
NÚrbÚcpuT)Úmap_locationÚweights_only)rê   )Ústrict)r   ÚstrÚosÚpathÚ
expanduserÚisfileÚopenr    ÚloadÚdevicer
   Úload_state_dictÚeval)rC   ræ   ÚlocalÚfÚpretrained_dicts   &&   r   râ   ÚDexiNed.load_from_fileý   s¯   € ô 2<¸IÄs×1KÒ1K”—‘×"Ñ" 9Ô-ÐQUˆØÒ¤§¡§¡°×!6Ò!6Ü�e˜T×"Ô" aÜ"'§*¢*¨Q¼U¿\º\È%Ó=PÐ_cÔ"d�÷ #Ð"ô 7°yÌuÏ|Ê|Ð\aÓObÔcˆOØ×Ñ˜_°TÐÔ:Ø�	‰	Ž÷ #×"ús   Á5.C2Ã2D	c               ó    € V ^8„  d   QhRRRR/# )r   rH   rI   r   rm   r   )r   s   "r   r   rÆ     s   € ÷ 6:ñ 6:˜lð 6:Ð/Añ 6:r   c                óÞ  € V P                  V4      pV P                  V4      pV P                  V4      pV P                  V4      pWS,           pV P	                  V4      pV P                  V4      pV P                  Wh.4      w  ršV P                  V	4      pW·,           pV P                  V4      pV P                  V4      pV P                  W¾,           4      pV P                  WÏ.4      w  pp
V P                  V4      pVV,           pV P                  V4      pV P                  V4      pV P                  VV.4      w  pp
VV,           pV P                  V4      pV P                  VV.4      w  pp
VP                   RR pV P#                  VV4      pV P%                  VV4      pV P'                  V	V4      pV P)                  VV4      pV P+                  VV4      pV P-                  VV4      pVVVVVV.# )a´  Compute the six multi-scale DexiNed side-output edge maps.

Args:
    x: Input BGR image tensor in :math:`[0, 255]` minus the BGR mean ``[103.939, 116.779, 123.68]``, with
        shape ``(B, 3, H, W)``, where :math:`B` is batch size, :math:`H` is height, and :math:`W` is
        width.

Returns:
    List of six tensors aligned to the input spatial size. Each tensor
    is a side-output edge response used by the final fusion block.
Néþÿÿÿ)rÉ   rÑ   rÊ   rÐ   rÒ   r×   rË   rÓ   rÖ   rØ   rÌ   rÔ   rÙ   rÍ   rÚ   rÎ   r%   rÛ   rÜ   rÝ   rÞ   rß   rà   ) rC   rH   rÉ   Úblock_1_siderÊ   Úblock_2_downÚblock_2_addÚblock_2_sideÚblock_3_pre_denseÚblock_3Ú_Úblock_3_downÚblock_3_addÚblock_3_sideÚblock_2_resize_halfÚblock_4_pre_denseÚblock_4Úblock_4_downÚblock_4_addÚblock_4_sideÚblock_5_pre_denseÚblock_5Úblock_5_addÚblock_6_pre_denseÚblock_6r¢   Úout_1Úout_2Úout_3Úout_4Úout_5Úout_6s    &&                              r   Úget_featuresÚDexiNed.get_features  sã  € ð —,‘,˜q“/ˆØ—{‘{ 7Ó+ˆð —,‘,˜wÓ'ˆØ—|‘| GÓ,ˆØ"Õ1ˆØ—{‘{ ;Ó/ˆð !×,Ñ,¨\Ó:ÐØ—]‘] KÐ#CÓD‰
ˆØ—|‘| GÓ,ˆØ"Õ1ˆØ—{‘{ ;Ó/ˆð #×.Ñ.¨|Ó<ÐØ ×,Ñ,¨\Õ-OÓPÐØ—]‘] KÐ#CÓD‰
ˆ�Ø—|‘| GÓ,ˆØ" \Õ1ˆØ—{‘{ ;Ó/ˆð !×,Ñ,¨\Ó:ÐØ—]‘] KÐ1BÐ#CÓD‰
ˆ�Ø Õ,ˆð !×,Ñ,¨WÓ5ÐØ—]‘] KÐ1BÐ#CÓD‰
ˆ�ð —G‘G˜B˜C�Lˆ	Ø—‘ ¨Ó3ˆØ—‘ ¨Ó3ˆØ—‘ ¨Ó3ˆØ—‘ ¨Ó3ˆØ—‘ ¨Ó3ˆØ—‘ ¨Ó3ˆØ�u˜e U¨E°5Ð9Ð9r   c               ó    € V ^8„  d   QhRRRR/# rG   r   )r   s   "r   r   rÆ   I  s   € ÷ )ñ )˜ð )¨,ñ )r   c                óv   € V P                  V4      p\        P                  ! V^R7      pV P                  V4      # )a!  Predict a fused edge map from a BGR image batch.

Args:
    x: Input BGR image tensor in :math:`[0, 255]` minus the BGR mean ``[103.939, 116.779, 123.68]``, with
        shape ``(B, 3, H, W)``.

Returns:
    Tensor with shape ``(B, 1, H, W)`` containing the fused DexiNed
    edge logits.
rL   )r  r    Úcatrá   )rC   rH   r�   rá   s   &&  r   rR   ÚDexiNed.forwardI  s5   € ð ×$Ñ$ QÓ'ˆô —I’I˜h¨AÔ.ˆ	Ø�~‰~˜iÓ(Ð(r   )rÉ   rÊ   rá   rË   rÌ   rÍ   rÎ   rÐ   rÖ   r×   rØ   rÙ   rÚ   rÑ   rÒ   rÓ   rÔ   rÕ   rÛ   rÜ   rÝ   rÞ   rß   rà   )éÿÿÿÿr•   r  r  )r  rK   r  r  )rU   rV   rW   rX   Ú__doc__rÂ   Ú__annotations__rÃ   r:   râ   r  rR   rY   rZ   r[   s   @r   r   r   ¿   sC   ø‡ ñò& 4CÐÐ0ÓBÚ4CÐÐ1ÓC÷%$ð %$õNõ(6:÷p)ó )r   )"Ú__conditional_annotations__Ú
__future__r   rî   Úcollectionsr   Útypingr   r   r    Útorch.nn.functionalr   Ú
functionalrN   Úkornia.core.checkr   Úkornia.core.downloadr	   r
   Úkornia.core.mixin.onnxr   Ú__all__r   r!  r+   ÚModuler-   r�   r]   ru   r†   rª   r¸   r   )r"  s   @r   Ú<module>r-     sÊ   øð÷( #Ð "ã 	Ý #ß %ã ß Ð Ý å *ß AÝ 2àˆ+€ñ ˆ9Ð,Ó-ØDð€€oó õ)ô01ˆr�y‰yô 1ô(%�"—-‘-ô %ô0�"—-‘-ô ô>�"—)‘)ô >ô>�b—i‘iô ô<�b—m‘mô <ô*Y)ˆo˜rŸy™yö Y)r   