§
    ·¯¸j­j  ã                   óR  — d Z ddlmZ ddlmZmZmZmZmZm	Z	 ddl
Z
ddlmZ ddlmZmZ ddlmZmZmZmZmZmZmZmZmZ ddlmZ dd	lmZ dd
lmZm Z  ddl!m"Z"m#Z# dgZ$ G d„ dej%        ¦  «        Z& G d„ dej%        ¦  «        Z' G d„ dej%        ¦  «        Z( G d„ dej%        ¦  «        Z) G d„ dej%        ¦  «        Z*d3d„Z+d4d„Z, e" e,d¬¦  «         e,d¬¦  «         e,d¬¦  «         e,d¬¦  «         e,d¬¦  «         e,d¬¦  «         e,ddddd¬¦  «         e,ddddd¬¦  «         e,ddddd¬¦  «         e,ddddd¬¦  «         e,dd¬ ¦  «         e,dd¬ ¦  «        d!œ¦  «        Z-d"e*fd#„Z.d5d%„Z/e#d5d&e*fd'„¦   «         Z0e#d5d&e*fd(„¦   «         Z1e#d5d&e*fd)„¦   «         Z2e#d5d&e*fd*„¦   «         Z3e#d5d&e*fd+„¦   «         Z4e#d5d&e*fd,„¦   «         Z5e#d5d&e*fd-„¦   «         Z6e#d5d&e*fd.„¦   «         Z7e#d5d&e*fd/„¦   «         Z8e#d5d&e*fd0„¦   «         Z9e#d5d&e*fd1„¦   «         Z:e#d5d&e*fd2„¦   «         Z;dS )6a[   FocalNet

As described in `Focal Modulation Networks` - https://arxiv.org/abs/2203.11926

Significant modifications and refactoring from the original impl at https://github.com/microsoft/FocalNet

This impl is/has:
* fully convolutional, NCHW tensor layout throughout, seemed to have minimal performance impact but more flexible
* re-ordered downsample / layer so that striding always at beginning of layer (stage)
* no input size constraints or input resolution/H/W tracking through the model
* torchscript fixed and a number of quirks cleaned up
* feature extraction support via `features_only=True`
é    )Úpartial)ÚCallableÚListÚOptionalÚTupleÚTypeÚUnionN©ÚIMAGENET_DEFAULT_MEANÚIMAGENET_DEFAULT_STD)	Úget_device_dtypeÚMlpÚDropPathÚLayerNorm2dÚLayerScale2dÚtrunc_normal_ÚClassifierHeadÚNormMlpClassifierHeadÚcalculate_drop_path_ratesé   )Úbuild_model_with_cfg)Úfeature_take_indices)Únamed_applyÚ
checkpoint)Úgenerate_default_cfgsÚregister_modelÚFocalNetc                   ór   ‡ — e Zd Zdddddeddfdededed	ed
ededededeej	                 fˆ fd„Z
d„ Zˆ xZS )ÚFocalModulationé   TFç        NÚdimÚfocal_windowÚfocal_levelÚfocal_factorÚbiasÚuse_post_normÚnormalize_modulatorÚ	proj_dropÚ
norm_layerc                 óª  •— |
|dœ}t          ¦   «                              ¦   «          || _        || _        || _        || _        || _        || _        ||| j        dz   g| _        t          j
        |d|z  | j        dz   z   fd|dœ|¤Ž| _        t          j
        ||fd|dœ|¤Ž| _        t          j        ¦   «         | _        t          j
        ||fddi|¤Ž| _        t          j        |¦  «        | _        t          j        ¦   «         | _        g | _        t+          | j        ¦  «        D ]…}| j        |z  | j        z   }| j                             t          j        t          j
        ||f|||dz  ddœ|¤Žt          j        ¦   «         ¦  «        ¦  «         | j                             |¦  «         Œ†| j        r	 |	|fi |¤Žnt          j        ¦   «         | _        d S )N©ÚdeviceÚdtyper   r    )Úkernel_sizer&   r/   F)r/   ÚgroupsÚpaddingr&   )ÚsuperÚ__init__r"   r#   r$   r%   r'   r(   Úinput_splitÚnnÚConv2dÚfÚhÚGELUÚactÚprojÚDropoutr)   Ú
ModuleListÚfocal_layersÚkernel_sizesÚrangeÚappendÚ
SequentialÚIdentityÚnorm)Úselfr"   r#   r$   r%   r&   r'   r(   r)   r*   r-   r.   ÚddÚkr/   Ú	__class__s                  €úd/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/timm/models/focalnet.pyr3   zFocalModulation.__init__/   sñ  ø€ ð ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐàˆŒØ(ˆÔØ&ˆÔØ(ˆÔØ*ˆÔØ#6ˆÔ Ø  dÔ&6¸Ñ&:Ð;ˆÔå”˜3  C¡¨4Ô+;¸aÑ+?Ñ @ÐaÈaÐVZÐaÐaÐ^`ÐaÐaˆŒÝ”˜3 ÐD°¸ÐDÐDÀÐDÐDˆŒå”7‘9”9ˆŒÝ”I˜c 3Ð<Ð<°AÐ<¸Ð<Ð<ˆŒ	Ýœ IÑ.Ô.ˆŒÝœM™OœOˆÔàˆÔÝ�tÔ'Ñ(Ô(ð 	2ð 	2ˆAØÔ+¨aÑ/°$Ô2CÑCˆKØÔ×$Ò$¥R¤]Ý”	˜#˜sÐt°ÀCÐQ\Ð`aÑQaÐhmÐtÐtÐqsÐtÐtÝ”‘	”	ñ&ô &ñ ô ð ð Ô×$Ò$ [Ñ1Ô1Ð1Ð1Ø-1Ô-?ÐR�J�J˜sÐ)Ð) bÐ)Ð)Ð)ÅRÄ[Á]Ä]ˆŒ	ˆ	ˆ	ó    c                 óF  — |                       |¦  «        }t          j        || j        d¦  «        \  }}}d}t	          | j        ¦  «        D ]'\  }} ||¦  «        }|||d d …||dz   …f         z  z   }Œ(|                      |                     dd¬¦  «        ¦  «        }|||d d …| j        d …f         z  z   }| j	        r|| j        dz   z  }||  
                    |¦  «        z  }	|                      |	¦  «        }	|                      |	¦  «        }	|                      |	¦  «        }	|	S )Nr   r   )r    é   T)Úkeepdim)r7   ÚtorchÚsplitr4   Ú	enumerater>   r:   Úmeanr$   r(   r8   rD   r;   r)   )
rE   ÚxÚqÚctxÚgatesÚctx_allÚlÚfocal_layerÚ
ctx_globalÚx_outs
             rI   ÚforwardzFocalModulation.forwardZ   s=  € à�FŠF�1‰IŒIˆÝœ A tÔ'7¸Ñ;Ô;‰ˆˆ3�ð ˆÝ'¨Ô(9Ñ:Ô:ð 	8ð 	8‰NˆAˆ{Ø�+˜cÑ"Ô"ˆCØ  e¨A¨A¨A¨q°°Q±¨w¨JÔ&7Ñ 7Ñ7ˆGˆGØ—X’X˜cŸhšh v°t˜hÑ<Ô<Ñ=Ô=ˆ
Ø˜J¨¨q¨q¨q°$Ô2BÐ2CÐ2CÐ/CÔ)DÑDÑDˆð Ô#ð 	7Ø Ô!1°AÑ!5Ñ6ˆGð �D—F’F˜7‘O”OÑ#ˆØ—	’	˜%Ñ Ô ˆð —	’	˜%Ñ Ô ˆØ—’˜uÑ%Ô%ˆØˆrJ   )Ú__name__Ú
__module__Ú__qualname__r   ÚintÚboolÚfloatr   r5   ÚModuler3   r[   Ú__classcell__©rH   s   @rI   r   r   .   sÚ   ø€ € € € € ð !"ØØ"'Ø(-Ø!Ø*5ØØð)Sð )Sàð)Sð ð)Sð ð	)Sð
 ð)Sð ð)Sð  ð)Sð "&ð)Sð ð)Sð ˜RœYœð)Sð )Sð )Sð )Sð )Sð )SðVð ð ð ð ð ð rJ   r   c                   ó¸   ‡ — e Zd ZdZdddddddddej        eddfd	ed
ededede	de	de	de
e         dededeej                 deej                 fˆ fd„Zd„ Zˆ xZS )ÚFocalNetBlockz% Focal Modulation Network Block.
    ç      @r   rL   Fç-Cëâ6?r!   Nr"   Ú	mlp_ratior$   r#   r'   Úuse_post_norm_in_modulationr(   Úlayerscale_valuer)   Ú	drop_pathÚ	act_layerr*   c           
      óŽ  •— ||dœ}t          ¦   «                              ¦   «          || _        || _        || _        || _        || _        |s	 ||fi |¤Žnt          j        ¦   «         | _	        t          |f|| j        |||	|dœ|¤Ž| _        |r	 ||fi |¤Žnt          j        ¦   «         | _        |�t          ||fi |¤Žnt          j        ¦   «         | _        |
dk    rt          |
¦  «        nt          j        ¦   «         | _        |s	 ||fi |¤Žnt          j        ¦   «         | _        t%          d|t'          ||z  ¦  «        ||	ddœ|¤Ž| _        |r	 ||fi |¤Žnt          j        ¦   «         | _        |�t          ||fi |¤Žnt          j        ¦   «         | _        |
dk    rt          |
¦  «        nt          j        ¦   «         | _        dS )ap  
        Args:
            dim: Number of input channels.
            mlp_ratio: Ratio of mlp hidden dim to embedding dim.
            focal_level: Number of focal levels.
            focal_window: Focal window size at first focal level.
            use_post_norm: Whether to use layer norm after modulation.
            use_post_norm_in_modulation: Whether to use layer norm in modulation.
            layerscale_value: Initial layerscale value.
            proj_drop: Dropout rate.
            drop_path: Stochastic depth rate.
            act_layer: Activation layer.
            norm_layer: Normalization layer.
        r,   )r#   r$   r'   r(   r)   r*   Nr!   T)Úin_featuresÚhidden_featuresrm   ÚdropÚuse_conv© )r2   r3   r"   ri   r#   r$   r'   r5   rC   Únorm1r   Ú
modulationÚ
norm1_postr   Úls1r   Ú
drop_path1Únorm2r   r_   ÚmlpÚ
norm2_postÚls2Ú
drop_path2)rE   r"   ri   r$   r#   r'   rj   r(   rk   r)   rl   rm   r*   r-   r.   rF   rH   s                   €rI   r3   zFocalNetBlock.__init__y   s  ø€ ð> ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐØˆŒØ"ˆŒà(ˆÔØ&ˆÔØ*ˆÔà2?ÐR�Z�Z Ð*Ð* rÐ*Ð*Ð*ÅRÄ[Á]Ä]ˆŒ
Ý)Øð	
à%ØÔ(Ø5Ø 3ØØ!ð	
ð 	
ð ð	
ð 	
ˆŒð 4AÐS˜*˜* SÐ/Ð/¨BÐ/Ð/Ð/ÅbÄkÁmÄmˆŒØ@PÐ@\•< Ð%5Ð<Ð<¸Ð<Ð<Ð<ÕbdÔbmÑboÔboˆŒØ1:¸R²°�( 9Ñ-Ô-Ð-ÅRÄ[Á]Ä]ˆŒà2?ÐR�Z�Z Ð*Ð* rÐ*Ð*Ð*ÅRÄ[Á]Ä]ˆŒ
Ýð 
ØÝ  i¡Ñ0Ô0ØØØð
ð 
ð ð
ð 
ˆŒð 4AÐS˜*˜* SÐ/Ð/¨BÐ/Ð/Ð/ÅbÄkÁmÄmˆŒØ@PÐ@\•< Ð%5Ð<Ð<¸Ð<Ð<Ð<ÕbdÔbmÑboÔboˆŒØ1:¸R²°�( 9Ñ-Ô-Ð-ÅRÄ[Á]Ä]ˆŒˆˆrJ   c                 ó¦  — |}|                       |¦  «        }|                      |¦  «        }|                      |¦  «        }||                      |                      |¦  «        ¦  «        z   }||                      |                      |                      |                      |  	                    |¦  «        ¦  «        ¦  «        ¦  «        ¦  «        z   }|S ©N)
rt   ru   rv   rx   rw   r}   r|   r{   rz   ry   )rE   rR   Úshortcuts      rI   r[   zFocalNetBlock.forward½   s¥   € Øˆð �JŠJ�q‰MŒMˆØ�OŠO˜AÑÔˆØ�OŠO˜AÑÔˆØ�t—’ t§x¢x°¡{¤{Ñ3Ô3Ñ3ˆð �—’ §¢¨¯ª¸¿ºÀ$Ç*Â*ÈQÁ-Ä-Ñ9PÔ9PÑ)QÔ)QÑ RÔ RÑSÔSÑSˆàˆrJ   )r\   r]   r^   Ú__doc__r5   r9   r   r_   ra   r`   r   r   rb   r3   r[   rc   rd   s   @rI   rf   rf   u   s3  ø€ € € € € ðð ð  "Ø Ø !Ø"'Ø05Ø(-Ø04Ø!Ø!Ø)+¬Ø*5ØØðBSð BSàðBSð ðBSð ð	BSð
 ðBSð  ðBSð *.ðBSð "&ðBSð ' uœoðBSð ðBSð ðBSð ˜BœI”ðBSð ˜RœYœðBSð BSð BSð BSð BSð BSðHð ð ð ð ð ð rJ   rf   c                    óî   ‡ — e Zd ZdZdddddddddddeddfd	ed
ededededededededededee         dede	ee
e         f         deej                 fˆ fd„Zej        j        dd„¦   «         Zd„ Zˆ xZS )ÚFocalNetStagez4 A basic Focal Transformer layer for one stage.
    rg   Tr   Frh   r!   Nr"   Úout_dimÚdepthri   Ú
downsampler$   r#   Úuse_overlap_downr'   rj   r(   rk   r)   rl   r*   c                 ór  •‡‡‡‡‡	‡
‡‡‡‡‡‡— ||dœŠt          ¦   «                              ¦   «          || _        || _        d| _        |rt          d|‰d|‰dœ‰¤Ž| _        nt          j        ¦   «         | _        t          j	        ˆˆˆˆˆˆˆˆˆˆˆ	ˆ
fd„t          |¦  «        D ¦   «         ¦  «        | _        dS )a8  
        Args:
            dim: Number of input channels.
            out_dim: Number of output channels.
            depth: Number of blocks.
            mlp_ratio: Ratio of mlp hidden dim to embedding dim.
            downsample: Downsample layer at start of the layer.
            focal_level: Number of focal levels
            focal_window: Focal window size at first focal level
            use_overlap_down: User overlapped convolution in downsample layer.
            use_post_norm: Whether to use layer norm after modulation.
            use_post_norm_in_modulation: Whether to use layer norm in modulation.
            layerscale_value: Initial layerscale value
            proj_drop: Dropout rate for projections.
            drop_path: Stochastic depth rate.
            norm_layer: Normalization layer.
        r,   Fr    )Úin_chsÚout_chsÚstrideÚoverlapr*   c                 óz   •— g | ]7}t          d‰
‰‰‰‰‰‰	‰‰t          ‰t          ¦  «        r‰|         n‰‰d œ‰¤Ž‘Œ8S ))r"   ri   r$   r#   r'   rj   r(   rk   r)   rl   r*   rs   )rf   Ú
isinstanceÚlist)Ú.0ÚirF   rl   r$   r#   rk   ri   r*   r(   r„   r)   r'   rj   s     €€€€€€€€€€€€rI   ú
<listcomp>z*FocalNetStage.__init__.<locals>.<listcomp>  s�   ø€ ð %#ð %#ð %#ð õ ð ØØ#Ø'Ø)Ø+Ø,GØ$7Ø!1Ø#Ý*4°YÅÑ*EÔ*EÐT˜) Aœ,˜,È9Ø%ðð ð ðð ð%#ð %#ð %#rJ   Nrs   )r2   r3   r"   r…   Úgrad_checkpointingÚ
Downsampler†   r5   rC   r=   r@   Úblocks)rE   r"   r„   r…   ri   r†   r$   r#   r‡   r'   rj   r(   rk   r)   rl   r*   r-   r.   rF   rH   s     ` ` `` ```````  @€rI   r3   zFocalNetStage.__init__Ð   s  øøøøøøøøøøøøø€ ðJ ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐØˆŒØˆŒ
Ø"'ˆÔàð 
	,Ý(ð ØØØØ(Ø%ðð ð ðð ˆDŒOˆOõ !œk™mœmˆDŒOõ ”mð %#ð %#ð %#ð %#ð %#ð %#ð %#ð %#ð %#ð %#ð %#ð %#ð %#ð %#ð %#õ ˜5‘\”\ð%#ñ %#ô %#ñ $ô $ˆŒˆˆrJ   c                 ó   — || _         d S r   )r“   )rE   Úenables     rI   Úset_grad_checkpointingz$FocalNetStage.set_grad_checkpointing  s   € à"(ˆÔÐÐrJ   c                 óÆ   — |                       |¦  «        }| j        D ]C}| j        r/t          j                             ¦   «         st          ||¦  «        }Œ8 ||¦  «        }ŒD|S r   )r†   r•   r“   rN   ÚjitÚis_scriptingr   )rE   rR   Úblks      rI   r[   zFocalNetStage.forward  si   € Ø�OŠO˜AÑÔˆØ”;ð 	ð 	ˆCØÔ&ð ­u¬y×/EÒ/EÑ/GÔ/Gð Ý˜s AÑ&Ô&��à�C˜‘F”F��ØˆrJ   ©T)r\   r]   r^   r�   r   r_   ra   r`   r   r	   r   r   r5   rb   r3   rN   rš   Úignorer˜   r[   rc   rd   s   @rI   rƒ   rƒ   Ì   sp  ø€ € € € € ðð ð  "Ø#Ø Ø !Ø%*Ø"'Ø05Ø(-Ø04Ø!Ø35Ø*5ØØð%G$ð G$àðG$ð ðG$ð ð	G$ð
 ðG$ð ðG$ð ðG$ð ðG$ð #ðG$ð  ðG$ð *.ðG$ð "&ðG$ð ' uœoðG$ð ðG$ð ˜U D¨¤KÐ/Ô0ðG$ð  ˜RœYœð!G$ð G$ð G$ð G$ð G$ð G$ðR „YÔð)ð )ð )ñ Ôð)ðð ð ð ð ð ð rJ   rƒ   c                   óh   ‡ — e Zd Z	 	 	 	 	 ddededededeeej                          f
ˆ fd	„Z	d
„ Z
ˆ xZS )r”   é   FNr‰   rŠ   r‹   rŒ   r*   c                 ó$  •— ||dœ}t          ¦   «                              ¦   «          || _        d}	|}
|r|dv sJ ‚|dk    rd\  }
}	n|dk    rd\  }
}	t          j        ||f|
||	dœ|¤Ž| _        |�	 ||fi |¤Žnt          j        ¦   «         | _        d	S )
a  

        Args:
            in_chs: Number of input image channels.
            out_chs: Number of linear projection output channels.
            stride: Downsample stride.
            overlap: Use overlapping convolutions if True.
            norm_layer: Normalization layer.
        r,   r   )r    r    r    )é   r    r    )rL   r   )r/   r‹   r1   N)r2   r3   r‹   r5   r6   r;   rC   rD   )rE   r‰   rŠ   r‹   rŒ   r*   r-   r.   rF   r1   r/   rH   s              €rI   r3   zDownsample.__init__)  sÍ   ø€ ð& ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐØˆŒØˆØˆØð 	,Ø˜VÐ#Ð#Ð#Ð#Ø˜Š{ˆ{Ø'+Ñ$�˜W˜WØ˜1’�Ø'+Ñ$�˜WÝ”I˜f gÐm¸;ÈvÐ_fÐmÐmÐjlÐmÐmˆŒ	Ø1;Ð1G�J�J˜wÐ-Ð-¨"Ð-Ð-Ð-ÍRÌ[É]Ì]ˆŒ	ˆ	ˆ	rJ   c                 óZ   — |                       |¦  «        }|                      |¦  «        }|S r   )r;   rD   ©rE   rR   s     rI   r[   zDownsample.forwardJ  s%   € Ø�IŠI�a‰LŒLˆØ�IŠI�a‰LŒLˆØˆrJ   )r    FNNN)r\   r]   r^   r_   r`   r   r   r5   rb   r3   r[   rc   rd   s   @rI   r”   r”   '  sª   ø€ € € € € ð Ø!Ø48ØØð[ð [àð[ð ð[ð ð	[ð
 ð[ð !  b¤i¤Ô1ð[ð [ð [ð [ð [ð [ðBð ð ð ð ð ð rJ   r”   c            (       óô  ‡ — e Zd ZdZdddddddd	d
d
d
d
dddddd eed¬¦  «        ddfdededededeedf         de	deedf         deedf         de
de
de
de
dee         de	d ee	         d!e	d"e	d#e	d$eej                 f&ˆ fd%„Zej        j        d&„ ¦   «         Zej        j        d=d'„¦   «         Zej        j        d>d)„¦   «         Zej        j        d*ej        fd+„¦   «         Zd?dedee         fd,„Z	 	 	 	 	 d@d.ej        d/eeeee         f                  d0e
d1e
d2ed3e
d*eeej                 eej        eej                 f         f         fd4„Z	 	 	 dAd/eeee         f         d6e
d7e
fd8„Zd9„ Zd=d:e
fd;„Zd<„ Zˆ xZ S )Br   z," Focal Modulation Networks (FocalNets)
    rL   éè  Úavgé`   ©r    r    é   r    rg   )r    r    r    r    ©rL   rL   rL   rL   FNç      ð?r!   gš™™™™™¹?gñhãˆµøä>)ÚepsÚin_chansÚnum_classesÚglobal_poolÚ	embed_dimÚdepths.ri   Úfocal_levelsÚfocal_windowsr‡   r'   rj   r(   Úhead_hidden_sizeÚhead_init_scalerk   Ú	drop_rateÚproj_drop_rateÚdrop_path_rater*   c                 óZ  •‡— t          ¦   «                              ¦   «          ||dœ}t          |¦  «        | _        ˆfd„t	          | j        ¦  «        D ¦   «         Š|| _        || _        ‰| _        ‰d         x| _        | _	        g | _
        t          d|‰d         |	|dœ|¤Ž| _        ‰d         }t          |t          |¦  «        ¦  «        }g }t	          | j        ¦  «        D ]£}‰|         }t          d||||         ||dk    ||         ||         |	|
|||||t          |d|…         ¦  «        t          |d|dz   …         ¦  «        …         |dœ|¤Ž}|}||gz  }| xj
        t!          |d	d
|z  z  d|› �¬¦  «        gz  c_
        Œ¤t#          j        |Ž | _        |r=t#          j        ¦   «         | _        || _	        t-          | j        |f||||dœ|¤Ž| _        n. || j        fi |¤Ž| _        t1          | j        |f||dœ|¤Ž| _        t3          t5          t6          |¬¦  «        | ¦  «         dS )aa  
        Args:
            in_chans: Number of input image channels.
            num_classes: Number of classes for classification head.
            embed_dim: Patch embedding dimension.
            depths: Depth of each Focal Transformer layer.
            mlp_ratio: Ratio of mlp hidden dim to embedding dim.
            focal_levels: How many focal levels at all stages. Note that this excludes the finest-grain level.
            focal_windows: The focal window size at all stages.
            use_overlap_down: Whether to use convolutional embedding.
            use_post_norm: Whether to use layernorm after modulation (it helps stabilize training of large models)
            layerscale_value: Value for layer scale.
            drop_rate: Dropout rate.
            drop_path_rate: Stochastic depth rate.
            norm_layer: Normalization layer.
        r,   c                 ó    •— g | ]
}‰d |z  z  ‘ŒS )r    rs   )r�   r‘   r±   s     €rI   r’   z%FocalNet.__init__.<locals>.<listcomp>  s"   ø€ ÐJÐJÐJ¨a�Y ! q¡&Ñ)ÐJÐJÐJrJ   éÿÿÿÿr   )r‰   rŠ   rŒ   r*   Nr   )r"   r„   r…   ri   r†   r$   r#   r‡   r'   rj   r(   rk   r)   rl   r*   r    r    úlayers.)Únum_chsÚ	reductionÚmodule)Úhidden_sizeÚ	pool_typer·   r*   )rÂ   r·   )r¶   rs   )r2   r3   ÚlenÚ
num_layersr@   r¯   r®   r±   Únum_featuresrµ   Úfeature_infor”   Ústemr   Úsumrƒ   Údictr5   rB   ÚlayersrC   rD   r   Úheadr   r   r   Ú_init_weights)rE   r®   r¯   r°   r±   r²   ri   r³   r´   r‡   r'   rj   r(   rµ   r¶   rk   r·   r¸   r¹   r*   r-   r.   rF   Úin_dimÚdprrÊ   Úi_layerr„   ÚlayerrH   s       `                        €rI   r3   zFocalNet.__init__T  sÎ  øø€ õP 	‰Œ×ÒÑÔÐØ¨Ð/Ð/ˆÝ˜f™+œ+ˆŒØJÐJÐJÐJµ5¸¼Ñ3IÔ3IÐJÑJÔJˆ	à&ˆÔØ ˆŒØ"ˆŒØ4=¸b´MÐAˆÔ˜DÔ1ØˆÔåð 
ØØ˜a”LØ$Ø!ð	
ð 
ð
 ð
ð 
ˆŒ	ð ˜1”ˆå'¨½¸F¹¼ÑDÔDˆØˆÝ˜Tœ_Ñ-Ô-ð 	qð 	qˆGØ Ô(ˆGÝ!ð ØØØ˜W”oØ#Ø" Qš;Ø(¨Ô1Ø*¨7Ô3Ø!1Ø+Ø,GØ$7Ø!1Ø(Ø�c &¨¨'¨Ô"2Ñ3Ô3µC¸¸|ÀÈ!Á¸|Ô8LÑ4MÔ4MÐMÔNØ%ðð ð  ð!ð ˆEð$ ˆFØ�u�gÑˆFØÐÔ¥$¨wÀ!ÀaÈ7ÁlÑBRÐ[nÐelÐ[nÐ[nÐ"oÑ"oÔ"oÐ!pÑpÐÔÐå”m VÐ,ˆŒàð 	Ýœ™œˆDŒIØ$4ˆDÔ!Ý-ØÔ!Øðð -Ø%Ø#Ø%ðð ð ðð ˆDŒIˆIð #˜
 4Ô#4Ð;Ð;¸Ð;Ð;ˆDŒIÝ&ØÔ!Øðð &Ø#ð	ð ð
 ðð ˆDŒIõ 	•G�M¸?ÐKÑKÔKÈTÑRÔRÐRÐRÐRrJ   c                 ó   — dhS )NÚ rs   ©rE   s    rI   Úno_weight_decayzFocalNet.no_weight_decayÄ  s	   € àˆtˆrJ   c                 ó4   — t          d|rddgng d¢¬¦  «        S )Nz^stem)z^layers\.(\d+)N©z^norm)iŸ† ))z^layers\.(\d+).downsample)r   )z^layers\.(\d+)\.\w+\.(\d+)NrÖ   )rÇ   r•   )rÉ   )rE   Úcoarses     rI   Úgroup_matcherzFocalNet.group_matcherÈ  sE   € åØð ðØ)Ø$ðð ðð ð ð

ñ 

ô 

ð 
	
rJ   Tc                 óT   — || _         | j        D ]}|                     |¬¦  «         Œd S )N)r—   )r“   rÊ   r˜   )rE   r—   rW   s      rI   r˜   zFocalNet.set_grad_checkpointingÖ  s?   € à"(ˆÔØ”ð 	4ð 	4ˆAØ×$Ò$¨FÐ$Ñ3Ô3Ð3Ð3ð	4ð 	4rJ   Úreturnc                 ó   — | j         j        S r   )rË   ÚfcrÓ   s    rI   Úget_classifierzFocalNet.get_classifierÜ  s   € àŒyŒ|ÐrJ   c                 ó\   — t          | ¦  «        }|| _         | j        j        |fd|i|¤Ž d S )NrÂ   )r   r¯   rË   Úreset)rE   r¯   r°   rF   s       rI   Úreset_classifierzFocalNet.reset_classifierà  s>   € Ý˜dÑ#Ô#ˆØ&ˆÔØˆŒ	Œ˜ÐAÐA¨{ÐA¸bÐAÐAÐAÐAÐArJ   ÚNCHWrR   ÚindicesrD   Ú
stop_earlyÚ
output_fmtÚintermediates_onlyc                 ó0  — |dv s
J d¦   «         ‚g }t          t          | j        ¦  «        |¦  «        \  }}	|                      |¦  «        }t          j                             ¦   «         s|s| j        }
n| j        d|	dz   …         }
t          | j        ¦  «        dz
  }t          |
¦  «        D ]I\  }} ||¦  «        }||v r5|r||k    r|                      |¦  «        }n|}| 	                    |¦  «         ŒJ|r|S ||k    r|                      |¦  «        }||fS )aï   Forward features that returns intermediates.

        Args:
            x: Input image tensor
            indices: Take last n blocks if int, all if None, select matching indices if sequence
            norm: Apply norm layer to compatible intermediates
            stop_early: Stop iterating over blocks when last desired intermediate hit
            output_fmt: Shape of intermediate feature outputs
            intermediates_only: Only return intermediate features
        Returns:

        )rá   zOutput shape must be NCHW.Nr   )
r   rÃ   rÊ   rÇ   rN   rš   r›   rP   rD   rA   )rE   rR   râ   rD   rã   rä   rå   ÚintermediatesÚtake_indicesÚ	max_indexÚstagesÚlast_idxÚfeat_idxÚstageÚx_inters                  rI   Úforward_intermediateszFocalNet.forward_intermediateså  sC  € ð* ˜YÐ&Ð&Ð&Ð(DÑ&Ô&Ð&ØˆÝ"6µs¸4¼;Ñ7GÔ7GÈÑ"QÔ"QÑˆ�ið �IŠI�a‰LŒLˆÝŒ9×!Ò!Ñ#Ô#ð 	1¨:ð 	1Ø”[ˆFˆFà”[  )¨a¡- Ô0ˆFå�t”{Ñ#Ô# aÑ'ˆÝ(¨Ñ0Ô0ð 	.ð 	.‰OˆH�eØ��a‘”ˆAØ˜<Ð'Ð'Øð  ˜H¨Ò0Ð0Ø"Ÿiši¨™lœl�G�Gà�GØ×$Ò$ WÑ-Ô-Ð-øàð 	!Ø Ð à�xÒÐØ—	’	˜!‘”ˆAà�-ÐÐrJ   r   Ú
prune_normÚ
prune_headc                 óâ   — t          t          | j        ¦  «        |¦  «        \  }}| j        d|dz   …         | _        |rt          j        ¦   «         | _        |r|                      dd¦  «         |S )z@ Prune layers not required for specified intermediates.
        Nr   r   rÒ   )r   rÃ   rÊ   r5   rC   rD   rà   )rE   râ   rð   rñ   rè   ré   s         rI   Úprune_intermediate_layersz"FocalNet.prune_intermediate_layers  sq   € õ #7µs¸4¼;Ñ7GÔ7GÈÑ"QÔ"QÑˆ�iØ”k . 9¨q¡= .Ô1ˆŒØð 	&Ýœ™œˆDŒIØð 	)Ø×!Ò! ! RÑ(Ô(Ð(ØÐrJ   c                 ó„   — |                       |¦  «        }|                      |¦  «        }|                      |¦  «        }|S r   )rÇ   rÊ   rD   r¤   s     rI   Úforward_featureszFocalNet.forward_features'  s4   € Ø�IŠI�a‰LŒLˆØ�KŠK˜‰NŒNˆØ�IŠI�a‰LŒLˆØˆrJ   Ú
pre_logitsc                 ó^   — |r|                       ||¬¦  «        n|                       |¦  «        S )N)rö   )rË   )rE   rR   rö   s      rI   Úforward_headzFocalNet.forward_head-  s-   € Ø6@ÐRˆt�yŠy˜ zˆyÑ2Ô2Ð2ÀdÇiÂiÐPQÁlÄlÐRrJ   c                 óZ   — |                       |¦  «        }|                      |¦  «        }|S r   )rõ   rø   r¤   s     rI   r[   zFocalNet.forward0  s-   € Ø×!Ò! !Ñ$Ô$ˆØ×Ò˜aÑ Ô ˆØˆrJ   ©Fr�   r   )NFFrá   F)r   FT)!r\   r]   r^   r�   r   r   r_   Ústrr   ra   r`   r   r   r5   rb   r3   rN   rš   rž   rÔ   rØ   r˜   rÝ   rà   ÚTensorr	   r   rï   ró   rõ   rø   r[   rc   rd   s   @rI   r   r   P  s¹  ø€ € € € € ðð ð
 Ø#Ø$ØØ&2Ø!Ø,8Ø-9Ø%*Ø"'Ø05Ø(-Ø.2Ø%(Ø04Ø!Ø$&Ø$'Ø*1¨'°+À4Ð*HÑ*HÔ*HØØð-nSð nSàðnSð ðnSð ð	nSð
 ðnSð ˜#˜s˜(”OðnSð ðnSð    S œ/ðnSð !  c œ?ðnSð #ðnSð  ðnSð *.ðnSð "&ðnSð ' sœmðnSð #ðnSð  ' uœoð!nSð" ð#nSð$ "ð%nSð& "ð'nSð( ˜RœYœð)nSð nSð nSð nSð nSð nSð` „YÔðð ñ Ôðð „YÔð
ð 
ð 
ñ Ôð
ð „YÔð4ð 4ð 4ñ Ôð4ð
 „YÔð ¤	ð ð ð ñ ÔððBð B¨Cð B¸hÀs¼mð Bð Bð Bð Bð 8<ØØ$Ø$Ø',ð0 ð 0 àŒ|ð0 ð ˜e C¨¨c¬ NÔ3Ô4ð0 ð ð	0 ð
 ð0 ð ð0 ð !%ð0 ð 
ˆt�E”LÔ! 5¨¬°t¸E¼LÔ7IÐ)IÔ#JÐJÔ	Kð0 ð 0 ð 0 ð 0 ðh ./Ø$Ø#ð	ð à˜3  S¤	˜>Ô*ðð ðð ð	ð ð ð ð ð ð ðSð S¨$ð Sð Sð Sð Sðð ð ð ð ð ð rJ   r¬   c                 ó  — t          | t          j        ¦  «        rEt          | j        d¬¦  «         | j        �&t          j                             | j        ¦  «         d S d S t          | t          j        ¦  «        r‡t          | j        d¬¦  «         | j        �$t          j                             | j        ¦  «         |rFd|v rD| j        j	         
                    |¦  «         | j        j	         
                    |¦  «         d S d S d S d S )Ng{®Gáz”?)Ústdúhead.fc)rŽ   r5   r6   r   Úweightr&   ÚinitÚzeros_ÚLinearÚdataÚmul_)rÀ   Únamer¶   s      rI   rÌ   rÌ   6  s  € Ý�&�"œ)Ñ$Ô$ð 
3Ý�f”m¨Ð-Ñ-Ô-Ð-ØŒ;Ð"ÝŒG�NŠN˜6œ;Ñ'Ô'Ð'Ð'Ð'ð #Ð"å	�F�BœIÑ	&Ô	&ð 3Ý�f”m¨Ð-Ñ-Ô-Ð-ØŒ;Ð"ÝŒG�NŠN˜6œ;Ñ'Ô'Ð'Øð 	3�I Ð%Ð%ØŒMÔ×#Ò# OÑ4Ô4Ð4ØŒKÔ×!Ò! /Ñ2Ô2Ð2Ð2Ð2ð3ð 3ð	3ð 	3Ð%Ð%rJ   rÒ   c                 ó6   — | dddddt           t          dddd	œ|¥S )
Nr¦   )rL   éà   r  )r¢   r¢   gÍÌÌÌÌÌì?Úbicubicz	stem.projrÿ   Úmit)Úurlr¯   Ú
input_sizeÚ	pool_sizeÚcrop_pctÚinterpolationrQ   rþ   Ú
first_convÚ
classifierÚlicenser
   )r  Úkwargss     rI   Ú_cfgr  D  s8   € àØ¨=ÀvØ¨Ý%Õ.BØ!°Øðð ð #ðð rJ   ztimm/)Ú	hf_hub_id)rL   é€  r  )é   r  iRU  )r  r  r  r  r¯   )r  r¯   )zfocalnet_tiny_srf.ms_in1kzfocalnet_small_srf.ms_in1kzfocalnet_base_srf.ms_in1kzfocalnet_tiny_lrf.ms_in1kzfocalnet_small_lrf.ms_in1kzfocalnet_base_lrf.ms_in1kzfocalnet_large_fl3.ms_in22kzfocalnet_large_fl4.ms_in22kzfocalnet_xlarge_fl3.ms_in22kzfocalnet_xlarge_fl4.ms_in22kzfocalnet_huge_fl3.ms_in22kzfocalnet_huge_fl4.ms_in22kÚmodelc                 ó¨  — |                       d| ¦  «        } d| v r| S dd l}i }|                     ¦   «         }|                      ¦   «         D �]\  }}|                     dd|¦  «        }|                     dd¦  «        }|                     dd	„ |¦  «        }d
|v r||vr|                     dd|¦  «        }|                     dd¦  «        }|                     dd¦  «        }||v rf||                              ¦   «         |                     ¦   «         k    r6||         j        |j        k    r |                     ||         j        ¦  «        }|||<   �Œ|S )Nr  zstem.proj.weightr   zgamma_([0-9])z
ls\1.gammaÚpatch_embedrÇ   zlayers.(\d+).downsamplec                 óT   — dt          |                      d¦  «        ¦  «        dz   › d�S )Nr½   r   z.downsample)r_   Úgroup)rR   s    rI   ú<lambda>z&checkpoint_filter_fn.<locals>.<lambda>|  s(   € Ð9cÅ3ÀqÇwÂwÈqÁzÄzÁ?Ä?ÐUVÑCVÐ9cÐ9cÐ9c€ rJ   rD   znorm([0-9])znorm\1_postzln.znorm.rË   rÿ   )	ÚgetÚreÚ
state_dictÚitemsÚsubÚreplaceÚnumelÚshapeÚreshape)r   r  r  Úout_dictÚ	dest_dictrG   Úvs          rI   Úcheckpoint_filter_fnr*  r  s^  € Ø—’ ¨Ñ4Ô4€JØ˜ZÐ'Ð'ØÐØ€I€I€IØ€HØ× Ò Ñ"Ô"€IØ× Ò Ñ"Ô"ð 
ñ 
‰ˆˆ1Ø�FŠFÐ# ]°AÑ6Ô6ˆØ�IŠI�m VÑ,Ô,ˆØ�FŠFÐ-Ð/cÐ/cÐefÑgÔgˆØ�Qˆ;ˆ;˜1 IÐ-Ð-Ø—’�~ ~°qÑ9Ô9ˆAØ�IŠI�e˜WÑ%Ô%ˆØ�IŠI�f˜iÑ(Ô(ˆØ�	ˆ>ˆ>˜i¨œl×0Ò0Ñ2Ô2°a·g²g±i´iÒ?Ð?ÀIÈaÄLÔDVÐZ[ÔZaÒDaÐDaØ—	’	˜) Aœ,Ô,Ñ-Ô-ˆAØˆ�‰‰Ø€OrJ   Fc           	      óü   — t          d„ t          |                     dd¦  «        ¦  «        D ¦   «         ¦  «        }|                     d|¦  «        }t	          t
          | |ft          t          d|¬¦  «        dœ|¤Ž}|S )Nc              3   ó    K  — | ]	\  }}|V — Œ
d S r   rs   )r�   r‘   Ú_s      rI   ú	<genexpr>z#_create_focalnet.<locals>.<genexpr>ˆ  s&   è è € Ð\Ð\¡d a¨ Ð\Ð\Ð\Ð\Ð\Ð\rJ   r²   )r   r   rL   r   Úout_indicesT)Úflatten_sequentialr/  )Úpretrained_filter_fnÚfeature_cfg)ÚtuplerP   r  Úpopr   r   r*  rÉ   )ÚvariantÚ
pretrainedr  Údefault_out_indicesr/  r  s         rI   Ú_create_focalnetr8  ‡  s�   € ÝÐ\Ð\­i¸¿
º
À8È\Ñ8ZÔ8ZÑ.[Ô.[Ð\Ñ\Ô\Ñ\Ô\ÐØ—*’*˜]Ð,?Ñ@Ô@€Kå Ý�'˜:ðå1Ý¨D¸kÐJÑJÔJðð ð ð	ð €Eð
 €LrJ   rÚ   c                 ó@   — t          dg d¢ddœ|¤Ž}t          dd| i|¤ŽS )Nr©   r¨   ©r²   r±   Úfocalnet_tiny_srfr6  rs   )r;  ©rÉ   r8  ©r6  r  Úmodel_kwargss      rI   r;  r;  “  s9   € åÐD˜|˜|˜|°rÐDÐD¸VÐDÐD€LÝÐWÐW¸JÐWÈ,ÐWÐWÐWrJ   c                 ó@   — t          dg d¢ddœ|¤Ž}t          dd| i|¤ŽS )N©r    r    é   r    r¨   r:  Úfocalnet_small_srfr6  rs   )rB  r<  r=  s      rI   rB  rB  ™  s9   € åÐE˜}˜}˜}¸ÐEÐE¸fÐEÐE€LÝÐXÐX¸ZÐXÈ<ÐXÐXÐXrJ   c                 ó@   — t          dg d¢ddœ|¤Ž}t          dd| i|¤ŽS )Nr@  é€   r:  Úfocalnet_base_srfr6  rs   )rE  r<  r=  s      rI   rE  rE  Ÿ  s9   € åÐF˜}˜}˜}¸ÐFÐF¸vÐFÐF€LÝÐWÐW¸JÐWÈ,ÐWÐWÐWrJ   c                 óF   — t          dg d¢dg d¢dœ|¤Ž}t          dd| i|¤ŽS )	Nr©   r¨   r«   ©r²   r±   r³   Úfocalnet_tiny_lrfr6  rs   )rH  r<  r=  s      rI   rH  rH  ¥  s@   € åÐ_˜|˜|˜|°rÈÈÈÐ_Ð_ÐX^Ð_Ð_€LÝÐWÐW¸JÐWÈ,ÐWÐWÐWrJ   c                 óF   — t          dg d¢dg d¢dœ|¤Ž}t          dd| i|¤ŽS )	Nr@  r¨   r«   rG  Úfocalnet_small_lrfr6  rs   )rJ  r<  r=  s      rI   rJ  rJ  «  s@   € åÐ`˜}˜}˜}¸ÈÈÈÐ`Ð`ÐY_Ð`Ð`€LÝÐXÐX¸ZÐXÈ<ÐXÐXÐXrJ   c                 óF   — t          dg d¢dg d¢dœ|¤Ž}t          dd| i|¤ŽS )	Nr@  rD  r«   rG  Úfocalnet_base_lrfr6  rs   )rL  r<  r=  s      rI   rL  rL  ±  s@   € åÐa˜}˜}˜}¸È,È,È,ÐaÐaÐZ`ÐaÐa€LÝÐWÐW¸JÐWÈ,ÐWÐWÐWrJ   c                 óV   — t          dg d¢dg d¢dgdz  ddddœ|¤Ž}t          dd
| i|¤ŽS )Nr@  éÀ   r«   é   r    Trh   ©r²   r±   r³   r´   r'   r‡   rk   Úfocalnet_large_fl3r6  rs   )rQ  r<  r=  s      rI   rQ  rQ  ¸  se   € åð TØˆ}ˆ}¨¸,¸,¸,ÐWXÐVYÐ\]ÑV]Ø¨TÀDðTð TàLRðTð T€Lõ ÐXÐX¸ZÐXÈ<ÐXÐXÐXrJ   c           
      óL   — t          d	g d¢dg d¢ddddœ|¤Ž}t          d
d| i|¤ŽS )Nr@  rN  ©r    r    r    r    Trh   ©r²   r±   r³   r'   r‡   rk   Úfocalnet_large_fl4r6  rs   )rU  r<  r=  s      rI   rU  rU  À  sY   € åð TØˆ}ˆ}¨¸,¸,¸,Ø¨TÀDðTð TàLRðTð T€Lõ ÐXÐX¸ZÐXÈ<ÐXÐXÐXrJ   c                 óV   — t          dg d¢dg d¢dgdz  ddddœ|¤Ž}t          dd
| i|¤ŽS )Nr@  é   r«   rO  r    Trh   rP  Úfocalnet_xlarge_fl3r6  rs   )rX  r<  r=  s      rI   rX  rX  È  se   € åð TØˆ}ˆ}¨¸,¸,¸,ÐWXÐVYÐ\]ÑV]Ø¨TÀDðTð TàLRðTð T€Lõ ÐYÐY¸jÐYÈLÐYÐYÐYrJ   c           
      óL   — t          d	g d¢dg d¢ddddœ|¤Ž}t          d
d| i|¤ŽS )Nr@  rW  rS  Trh   rT  Úfocalnet_xlarge_fl4r6  rs   )rZ  r<  r=  s      rI   rZ  rZ  Ð  sY   € åð TØˆ}ˆ}¨¸,¸,¸,Ø¨TÀDðTð TàLRðTð T€Lõ ÐYÐY¸jÐYÈLÐYÐYÐYrJ   c                 óX   — t          dg d¢dg d¢dgdz  dddddœ|¤Ž}t          dd
| i|¤ŽS )Nr@  é`  r«   rL   r    Trh   )r²   r±   r³   r´   r'   rj   r‡   rk   Úfocalnet_huge_fl3r6  rs   )r]  r<  r=  s      rI   r]  r]  Ø  sh   € åð vØˆ}ˆ}¨¸,¸,¸,ÐWXÐVYÐ\]ÑV]Ø¸ÈtÐfjðvð vàntðvð v€Lõ ÐWÐW¸JÐWÈ,ÐWÐWÐWrJ   c                 óN   — t          d	g d¢dg d¢dddddœ|¤Ž}t          d
d| i|¤ŽS )Nr@  r\  rS  Trh   )r²   r±   r³   r'   rj   r‡   rk   Úfocalnet_huge_fl4r6  rs   )r_  r<  r=  s      rI   r_  r_  à  s\   € åð vØˆ}ˆ}¨¸,¸,¸,Ø¸ÈtÐfjðvð vàntðvð v€Lõ ÐWÐW¸JÐWÈ,ÐWÐWÐWrJ   )Nr¬   )rÒ   rú   )<r�   Ú	functoolsr   Útypingr   r   r   r   r   r	   rN   Útorch.nnr5   Ú	timm.datar   r   Útimm.layersr   r   r   r   r   r   r   r   r   Ú_builderr   Ú	_featuresr   Ú_manipulater   r   Ú	_registryr   r   Ú__all__rb   r   rf   rƒ   r”   r   rÌ   r  Údefault_cfgsr*  r8  r;  rB  rE  rH  rJ  rL  rQ  rU  rX  rZ  r]  r_  rs   rJ   rI   ú<module>rk     s†  ððð ð& Ð Ð Ð Ð Ð Ø ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?à €€€Ø Ð Ð Ð Ð Ð à AÐ AÐ AÐ AÐ AÐ AÐ AÐ Að
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð 
ð +Ð *Ð *Ð *Ð *Ð *Ø +Ð +Ð +Ð +Ð +Ð +Ø 0Ð 0Ð 0Ð 0Ð 0Ð 0Ð 0Ð 0Ø <Ð <Ð <Ð <Ð <Ð <Ð <Ð <àˆ,€ðDð Dð Dð Dð D�b”iñ Dô Dð DðNTð Tð Tð Tð T�B”Iñ Tô Tð TðnXð Xð Xð Xð X�B”Iñ Xô Xð Xðv&ð &ð &ð &ð &�”ñ &ô &ð &ðRcð cð cð cð cˆrŒyñ cô cð cðL3ð 3ð 3ð 3ðð ð ð ð %Ð$Ø!% Øð"ñ "ô "à"& $Øð#ñ #ô #à!% Øð"ñ "ô "à!% Øð"ñ "ô "à"& $Øð#ñ #ô #à!% Øð"ñ "ô "ð $( 4ØØ ¨H¸sÐPUð$Wñ $Wô $Wð $( 4ØØ ¨H¸sÐPUð$Wñ $Wô $Wð %) DØØ ¨H¸sÐPUð%Wñ %Wô %Wð %) DØØ ¨H¸sÐPUð%Wñ %Wô %Wð #' $ØØð#ñ #ô #ð #' $ØØð#ñ #ô #ð; &ð  &ñ  ô  €ðF¨Hð ð ð ð ð*	ð 	ð 	ð 	ð ðXð X°Xð Xð Xð Xñ „ðXð
 ðYð Y°hð Yð Yð Yñ „ðYð
 ðXð X°Xð Xð Xð Xñ „ðXð
 ðXð X°Xð Xð Xð Xñ „ðXð
 ðYð Y°hð Yð Yð Yñ „ðYð
 ðXð X°Xð Xð Xð Xñ „ðXð ðYð Y°hð Yð Yð Yñ „ðYð ðYð Y°hð Yð Yð Yñ „ðYð ðZð Z°xð Zð Zð Zñ „ðZð ðZð Z°xð Zð Zð Zñ „ðZð ðXð X°Xð Xð Xð Xñ „ðXð ðXð X°Xð Xð Xð Xñ „ðXð Xð XrJ   