Ë
    éÓ^j¦– ã                   óê!  — d Z ddlZddlmZ ddlmZ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c 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mZm Z m!Z! ddl"m#Z# dd	l$m%Z% dd
l&m'Z' ddl(m)Z)m*Z*m+Z+ g d¢Z,�dde-de-de-de-fd„Z. G d„ dej^                  «      Z0 G d„ dej^                  «      Z1	 	 	 	 	 	 �dde-de-de-de-de-dee-   dee
ej^                        dej^                  fd„Z2	 	 	 	 	 	 �dde-de-de-de-de-dee-   dee
ej^                        dej^                  fd„Z3�dde4deee      fd„Z5	 	 	 	 	 	 	 	 �dde	ee
e0   e
e1   f   df   d e	e-df   d!e	e-df   d"e-d#e-d$e-d%e-d&e6d'e4d(e4de	ee	e7ej^                  f      eee7ef      f   fd)„Z8 G d*„ d+ej^                  «      Z9�dd,e7d-e6de9fd.„Z:�dd/e7dee7ef   fd0„Z;�dd/e7dee7ef   fd1„Z<�dd/e7dee7ef   fd2„Z=�dd/e7dee7ef   fd3„Z>�dd/e7dee7ef   fd4„Z?�dd/e7dee7ef   fd5„Z@ e*i d6 e=d7d8d9d:d;d<d=¬>«      “d? e=d7d@d9d:d;d<d=¬>«      “dA e>d7dB¬C«      “dD e>d7dE¬C«      “dF e?d7dG¬C«      “dH e=d7dId=¬J«      “dK e>d7dLdLdMd=¬N«      “dO e>d7dP¬C«      “dQ e>d7dR¬C«      “dS e?d7dTd;¬U«      “dV e=d7dW¬C«      “dX e>d7dLdLdM¬Y«      “dZ e=d7d[d=¬J«      “d\ e=d7d]¬C«      “d^ e=d7d_d=¬J«      “d` e=d7dad=dbdcdddedf¬g«      “dh e>d7di¬C«      “i dj e>d7dkd9d:dMd<df¬l«      “dm e>d7dn¬C«      “do e?d7dp¬C«      “dq e>d7dr¬C«      “ds e>d7dt¬C«      “du e>d7dv¬C«      “dw e>d7dx¬C«      “dy e>d7dz¬C«      “d{ e=d7d|¬C«      “d} e<d7d~¬C«      “d e=d7d€¬C«      “d� e=d7d‚¬C«      “dƒ e=d7d„d=¬J«      “d… e>d7dLdLd;d†dfd=¬‡«      “dˆ e>d7d‰d=¬J«      “dŠ e>d7d‹d=¬J«      “dŒ e?d7d�d=¬J«      “¥i dŽ e=d=¬�«      “d� e>d7d‘¬C«      “d’ e>d7d“¬C«      “d” e>d7d•¬C«      “d– e?d7d—¬C«      “d˜ e=d7d™d=dbdcd;dfde¬š«      “d› e>d7dœ¬C«      “d� e>d7dž¬C«      “dŸ e>d7d ¬C«      “d¡ e?d7d¢¬C«      “d£ e=d7d¤d=dbdcd;dfde¬š«      “d¥ e=«       “d¦ e=d7d§d=dbdcd;dfde¬š«      “d¨ e=d7d©¬C«      “dª e;d7d«d¬d­¬®«      “d¯ e;d7d°d¬d­¬®«      “d± e;d7d²d¬d­¬®«      “¥i d³ e;d7d´d9d:d<dµd¬d­¬¶«      “d· e;d7d¸d¬d­¬®«      “d¹ e;d7dºd9d:d<dµd¬d­¬¶«      “d» e;d7d¼d¬d­¬®«      “d½ e;d7d¾d9d:d<dµd¬d­¬¶«      “d¿ e;d7dÀd¬d­¬®«      “dÁ e;d7dÂd9d:d<dµd¬d­¬¶«      “dÃ e;d7dÄd¬d­¬®«      “dÅ e;d7dÆd9d:d<dµd¬d­¬¶«      “dÇ e=d7dÈdd¬U«      “dÉ e>d7dÊ¬C«      “dË e>d7dÌ¬C«      “dÍ e>d7dÎ¬C«      “dÏ e?d7dÐ¬C«      “dÑ e=d7dÒ¬C«      “dÓ e=d7dÔd=¬J«      “dÕ e=«       “¥i dÖ e>d7d×¬C«      “dØ e;d7dÙd¬d­¬®«      “dÚ e;d7dÛd¬d­¬®«      “dÜ e;d7dÝd¬d­¬®«      “dÞ e;d7dßd9d:d<dµd¬d­¬¶«      “dà e;d7dád9d:d<dµd¬d­¬¶«      “dâ e;d7dãdädå¬®«      “dæ e;d7dçdädå¬®«      “dè e;d7dédädå¬®«      “dê e;d7dëdädå¬®«      “dì e;d7dídädî¬®«      “dï e;d7dðdädî¬®«      “dñ e;d7dòdädî¬®«      “dó e;d7dôdädî¬®«      “dõ e;d7dödädî¬®«      “d÷ e;d7dødädî¬®«      “dù e;d7dúdädî¬®«      “¥i dû e;d7düdädî¬®«      “dý e;d7dþdädî¬®«      “dÿ e;d7�d dädî¬®«      “�d e;d7�ddädî¬®«      “�d e;d7�ddädî¬®«      “�d e=d7�dd=dbdcd;de�¬«      “�d e<d7�d	d;d†�¬
«      “�d e<d7�dd=d;d†�¬«      “�d e<d7�dd=d;d†�¬«      “�d e<d7�dd=dbdcd;de�¬«      “�d e>d7�dd=¬J«      “�d e>d7�dd=¬J«      “�d e?d7�dd=¬J«      “�d e<d7�dd=d;d†�¬«      “�d e<d7�dd=d;d†�¬«      “�d e=d=dbd;dc�¬«      “�d e=d7�dd=de�d d;df�d!¬š«      “¥i �d" e<d=¬�«      “�d# e<d=¬�«      “�d$ e=«       “�d% e=«       “�d& e>d7�d'd;¬U«      “�d( e>d7�d)d;¬U«      “�d* e?d7�d+d;¬U«      “�d, e=d7�d-¬C«      “�d. e=d=¬�«      “�d/ e=«       “�d0 e=«       “�d1 e=d7�d2d=dbdcd;dfde¬š«      “�d3 e=d=dbdc�¬4«      “�d5 e=d=dbdc�¬4«      “�d6 e=d7�d7d=¬J«      “�d8 e=d7�d9d=¬J«      “�d: e=d7�d;¬C«      “¥i �d< e=«       “�d= e>d7�d>¬C«      “�d? e>d7�d@d=¬J«      “�dA e=d7d=d;df�¬B«      “�dC e=d7d=d;df�¬B«      “�dD e=d7d;d†�dEdedfd=�¬F«      “�dG e=d7d=df�¬H«      “�dI e;d7�dJd=�dK�dLdf�¬M«      “�dN e;d7�dO�dJd=d;de�d �dLdf�¬P«	      “�dQ e=d7�dOd=d;df�¬R«      “�dS e=d7�dOd=d;df�¬R«      “�dT e=d7�dOd=d;df�¬R«      “�dU e=d7�dOd=d;df�¬R«      “�dV e=«       “�dW e=d7�dX¬C«      “�dY e=d=¬�«      “�dZ e=d=¬�«      “¥i �d[ e=d=¬�«      “�d\ e>d7�d]¬C«      “�d^ e=d=¬�«      “�d_ e>d7�d`d=¬J«      “�da e;d7�db�dc�dd�ded<�dJd=�¬f«      “�dg e;d7�dh�did:ddd†�dJd=�¬f«      “�dj e;d7�dkdbdcdfde�dJd=�¬f«      “�dl e;d7�dmdbdcdfde�dJd=�¬f«      “�dn e;d7�dodbdcdf�d!�dJd=�¬f«      “�dp e;d7�dqd†�dEdf�dL�dJd=�¬f«      “�dr e;d7�dsde�d df�dt�dJd=�¬f«      “�du e@d7�dv¬C«      “�dw e@d7�dx¬C«      “�dy e@d7�dz¬C«      “�d{ e@d7�d|¬C«      “�d} e@d7�d~¬C«      “�d e@d7�d€d=¬J«      “¥i �d� e@d7�d‚d=¬J«      “�dƒ e@d7�d„d=¬J«      “�d… e@d7�d†d=¬J«      “�d‡ e@d7�dˆd=¬J«      “�d‰ e@d7�dŠd=¬J«      “�d‹ e@d7�dŒd=¬J«      “�d� e@d7�dŽd=¬J«      “�d� e@d7�d�d=¬J«      “�d‘ e@d7�d’¬C«      “�d“ e@d7�d”¬C«      “�d• e@d7�d–¬C«      “�d— e@d7�d˜¬C«      “�d™ e@d7�dš¬C«      “�d› e@d7�dœ¬C«      “�d� e@d7�džd=¬J«      “�dŸ e;d7dLdLd;�dc�ddd=�¬ «      “¥«      ZAe)�dd-e6de9f�d¡„«       ZBe)�dd-e6de9f�d¢„«       ZCe)�dd-e6de9f�d£„«       ZDe)�dd-e6de9f�d¤„«       ZEe)�dd-e6de9f�d¥„«       ZFe)�dd-e6de9f�d¦„«       ZGe)�dd-e6de9f�d§„«       ZHe)�dd-e6de9f�d¨„«       ZIe)�dd-e6de9f�d©„«       ZJe)�dd-e6de9f�dª„«       ZKe)�dd-e6de9f�d«„«       ZLe)�dd-e6de9f�d¬„«       ZMe)�dd-e6de9f�d­„«       ZNe)�dd-e6de9f�d®„«       ZOe)�dd-e6de9f�d¯„«       ZPe)�dd-e6de9f�d°„«       ZQe)�dd-e6de9f�d±„«       ZRe)�dd-e6de9f�d²„«       ZSe)�dd-e6de9f�d³„«       ZTe)�dd-e6de9f�d´„«       ZUe)�dd-e6de9f�dµ„«       ZVe)�dd-e6de9f�d¶„«       ZWe)�dd-e6de9f�d·„«       ZXe)�dd-e6de9f�d¸„«       ZYe)�dd-e6de9f�d¹„«       ZZe)�dd-e6de9f�dº„«       Z[e)�dd-e6de9f�d»„«       Z\e)�dd-e6de9f�d¼„«       Z]e)�dd-e6de9f�d½„«       Z^e)�dd-e6de9f�d¾„«       Z_e)�dd-e6de9f�d¿„«       Z`e)�dd-e6de9f�dÀ„«       Zae)�dd-e6de9f�dÁ„«       Zbe)�dd-e6de9f�dÂ„«       Zce)�dd-e6de9f�dÃ„«       Zde)�dd-e6de9f�dÄ„«       Zee)�dd-e6de9f�dÅ„«       Zfe)�dd-e6de9f�dÆ„«       Zge)�dd-e6de9f�dÇ„«       Zhe)�dd-e6de9f�dÈ„«       Zie)�dd-e6de9f�dÉ„«       Zje)�dd-e6de9f�dÊ„«       Zke)�dd-e6de9f�dË„«       Zle)�dd-e6de9f�dÌ„«       Zme)�dd-e6de9f�dÍ„«       Zne)�dd-e6de9f�dÎ„«       Zoe)�dd-e6de9f�dÏ„«       Zpe)�dd-e6de9f�dÐ„«       Zqe)�dd-e6de9f�dÑ„«       Zre)�dd-e6de9f�dÒ„«       Zse)�dd-e6de9f�dÓ„«       Zte)�dd-e6de9f�dÔ„«       Zue)�dd-e6de9f�dÕ„«       Zve)�dd-e6de9f�dÖ„«       Zwe)�dd-e6de9f�d×„«       Zxe)�dd-e6de9f�dØ„«       Zye)�dd-e6de9f�dÙ„«       Zze)�dd-e6de9f�dÚ„«       Z{e)�dd-e6de9f�dÛ„«       Z|e)�dd-e6de9f�dÜ„«       Z}e)�dd-e6de9f�dÝ„«       Z~e)�dd-e6de9f�dÞ„«       Ze)�dd-e6de9f�dß„«       Z€e)�dd-e6de9f�dà„«       Z�e)�dd-e6de9f�dá„«       Z‚e)�dd-e6de9f�dâ„«       Zƒe)�dd-e6de9f�dã„«       Z„e)�dd-e6de9f�dä„«       Z…e)�dd-e6de9f�då„«       Z†e)�dd-e6de9f�dæ„«       Z‡e)�dd-e6de9f�dç„«       Zˆe)�dd-e6de9f�dè„«       Z‰e)�dd-e6f�dé„«       ZŠe)�dd-e6de9f�dê„«       Z‹e)�dd-e6de9f�dë„«       ZŒe)�dd-e6de9f�dì„«       Z�e)�dd-e6de9f�dí„«       ZŽe)�dd-e6de9f�dî„«       Z�e)�dd-e6de9f�dï„«       Z�e)�dd-e6de9f�dð„«       Z‘e)�dd-e6de9f�dñ„«       Z’ e+e“i �dòd¯“�dód±“�dôd·“�dõd»“�dödØ“�d÷dâ“�dødâ“�dùdâ“�dúdâ“�dûdì“�düdï“�dýdñ“�dþdó“�dÿdõ“�d d÷“�ddù“�ddû“i �ddý“�ddÿ“�d�d“�d�d“�d�du“�d�dw“�d	�dy“�d
�d{“�d�d}“�d�d“�d�d�“�d�dƒ“�d�d…“�d�d‡“�d�d‰“�d�d‹“�d�d�“¥�d��d‘�d“�d•�d—�d™�d›�d��d�dœ	¥«       y(  a*  PyTorch ResNet

This started as a copy of https://github.com/pytorch/vision 'resnet.py' (BSD-3-Clause) with
additional dropout and dynamic global avg/max pool.

ResNeXt, SE-ResNeXt, SENet, and MXNet Gluon stem/downsample variants, tiered stems added by Ross Wightman

Copyright 2019, Ross Wightman
é    N)Úpartial)ÚAnyÚDictÚListÚOptionalÚTupleÚTypeÚUnion©ÚIMAGENET_DEFAULT_MEANÚIMAGENET_DEFAULT_STD)ÚDropBlock2dÚDropPathÚAvgPool2dSameÚ
BlurPool2dÚ	LayerTypeÚcreate_attnÚget_attnÚget_act_layerÚget_norm_layerÚcreate_classifierÚ	create_aaÚ	to_ntupleé   )Úbuild_model_with_cfg)Úfeature_take_indices)Úcheckpoint_seq)Úregister_modelÚgenerate_default_cfgsÚregister_model_deprecations)ÚResNetÚ
BasicBlockÚ
BottleneckÚkernel_sizeÚstrideÚdilationÚreturnc                 ó(   — |dz
  || dz
  z  z   dz  }|S )Nr   é   © )r$   r%   r&   Úpaddings       ú]/var/www/ramen.bs-engineer-server.com/venv/lib/python3.12/site-packages/timm/models/resnet.pyÚget_paddingr-      s#   € Ø˜‘
˜h¨+¸©/Ñ:Ñ:¸qÑ@€GØ€Nó    c            !       ó¾  ‡ — e Zd ZdZdZdddddddej                  ej                  ddddddfdededede	ej                     d	ed
ededede	e   deej                     deej                     de	eej                        de	eej                        de	eej                        de	ej                     ddf ˆ fd„Zdd„Zdej                  dej                  fd„Zˆ xZS )r"   zoBasic residual block for ResNet.

    This is the standard residual block used in ResNet-18 and ResNet-34.
    r   Né@   ÚinplanesÚplanesr%   Ú
downsampleÚcardinalityÚ
base_widthÚreduce_firstr&   Úfirst_dilationÚ	act_layerÚ
norm_layerÚ
attn_layerÚaa_layerÚ
drop_blockÚ	drop_pathr'   c           	      ó|  •— ||dœ}t         ‰| �  «        |dk(  sJ d«       ‚|dk(  sJ d«       ‚||z  }|| j                  z  }|	xs |}	|duxr |dk(  xs |	|k7  }t        j                  ||fd|rdn||	|	d	d
œ|¤Ž| _         ||fi |¤Ž| _        |� |«       nt        j                  «       | _         |
d¬«      | _	        t        |f|||dœ|¤Ž| _        t        j                  ||fd||d	dœ|¤Ž| _         ||fi |¤Ž| _        t        ||fi |¤Ž| _         |
d¬«      | _        || _        || _        || _        || _        y)á¨  
        Args:
            inplanes: Input channel dimensionality.
            planes: Used to determine output channel dimensionalities.
            stride: Stride used in convolution layers.
            downsample: Optional downsample layer for residual path.
            cardinality: Number of convolution groups.
            base_width: Base width used to determine output channel dimensionality.
            reduce_first: Reduction factor for first convolution output width of residual blocks.
            dilation: Dilation rate for convolution layers.
            first_dilation: Dilation rate for first convolution layer.
            act_layer: Activation layer class.
            norm_layer: Normalization layer class.
            attn_layer: Attention layer class.
            aa_layer: Anti-aliasing layer class.
            drop_block: DropBlock layer class.
            drop_path: Optional DropPath layer instance.
        ©ÚdeviceÚdtyper   z)BasicBlock only supports cardinality of 1r0   z/BasicBlock does not support changing base widthNr)   é   F)r$   r%   r+   r&   ÚbiasT©Úinplace©Úchannelsr%   Úenable)r$   r+   r&   rD   )ÚsuperÚ__init__Ú	expansionÚnnÚConv2dÚconv1Úbn1ÚIdentityr<   Úact1r   ÚaaÚconv2Úbn2r   ÚseÚact2r3   r%   r&   r=   )Úselfr1   r2   r%   r3   r4   r5   r6   r&   r7   r8   r9   r:   r;   r<   r=   rA   rB   ÚddÚfirst_planesÚ	outplanesÚuse_aaÚ	__class__s                         €r,   rK   zBasicBlock.__init__)   sŽ  ø€ ðL ¨Ñ/ˆÜ‰ÑÔà˜aÒÐLÐ!LÓLÐØ˜RÒÐRÐ!RÓRÐØ Ñ-ˆØ˜TŸ^™^Ñ+ˆ	Ø'Ò3¨8ˆØ Ð%ÒU¨6°Q©;Ò+T¸.ÈHÑ:Tˆä—Y‘YØØð	
ð Ù‘1 FØ"Ø#Øñ	
ð ñ	
ˆŒ
ñ ˜lÑ1¨bÑ1ˆŒØ*4Ð*@™*œ,ÄbÇkÁkÃmˆŒÙ dÔ+ˆŒ	Ü˜HÐ`¨|ÀFÐSYÑ`Ð]_Ñ`ˆŒä—Y‘YØØð
ð ØØØñ
ð ñ
ˆŒ
ñ ˜iÑ.¨2Ñ.ˆŒä˜j¨)Ñ:°rÑ:ˆŒá dÔ+ˆŒ	Ø$ˆŒØˆŒØ ˆŒØ"ˆ�r.   c                 óš   — t        | j                  dd«      �4t        j                  j	                  | j                  j
                  «       yy©zLInitialize the last batch norm layer weights to zero for better convergence.ÚweightN)ÚgetattrrU   rM   ÚinitÚzeros_r`   ©rX   s    r,   Úzero_init_lastzBasicBlock.zero_init_last{   ó2   € ä�4—8‘8˜X tÓ,Ð8Ü�G‰G�N‰N˜4Ÿ8™8Ÿ?™?Õ+ð 9r.   Úxc                 óÒ  — |}| j                  |«      }| j                  |«      }| j                  |«      }| j                  |«      }| j	                  |«      }| j                  |«      }| j                  |«      }| j                  �| j                  |«      }| j                  �| j                  |«      }| j                  �| j                  |«      }||z  }| j                  |«      }|S ©N)rO   rP   r<   rR   rS   rT   rU   rV   r=   r3   rW   ©rX   rg   Úshortcuts      r,   ÚforwardzBasicBlock.forward€   sÅ   € Øˆà�J‰J�q‹MˆØ�H‰H�Q‹KˆØ�O‰O˜AÓˆØ�I‰I�a‹LˆØ�G‰G�A‹Jˆà�J‰J�q‹MˆØ�H‰H�Q‹Kˆà�7‰7ÐØ—‘˜“
ˆAà�>‰>Ð%Ø—‘˜qÓ!ˆAà�?‰?Ð&Ø—‘ xÓ0ˆHØ	ˆX‰ˆØ�I‰I�a‹Lˆàˆr.   ©r'   N©Ú__name__Ú
__module__Ú__qualname__Ú__doc__rL   rM   ÚReLUÚBatchNorm2dÚintr   ÚModuler	   rK   re   ÚtorchÚTensorrl   Ú__classcell__©r]   s   @r,   r"   r"   "   sh  ø„ ñð €Ið Ø.2Ø Ø Ø !ØØ,0Ø)+¯©Ø*,¯.©.Ø48Ø26Ø48Ø-1ØØñ%P#àðP#ð ðP#ð ð	P#ð
 ! §¡Ñ+ðP#ð ðP#ð ðP#ð ðP#ð ðP#ð % S™MðP#ð ˜BŸI™I‘ðP#ð ˜RŸY™Y™ðP#ð !  b§i¡i¡Ñ1ðP#ð ˜t B§I¡I™Ñ/ðP#ð !  b§i¡i¡Ñ1ðP#ð    §	¡	Ñ*ð!P#ð& 
õ'P#ód,ð
˜Ÿ™ð ¨%¯,©,÷ r.   r"   c            !       ó¾  ‡ — e Zd ZdZdZdddddddej                  ej                  ddddddfdededed	e	ej                     d
edededede	e   deej                     deej                     de	eej                        de	eej                        de	eej                        de	ej                     ddf ˆ fd„Zdd„Zdej                  dej                  fd„Zˆ xZS )r#   z{Bottleneck residual block for ResNet.

    This is the bottleneck block used in ResNet-50, ResNet-101, and ResNet-152.
    é   r   Nr0   r1   r2   r%   r3   r4   r5   r6   r&   r7   r8   r9   r:   r;   r<   r=   r'   c           
      ó  •— ||dœ}t         ‰| �  «        t        t        j                  ||dz  z  «      |z  «      }||z  }|| j
                  z  }|	xs |}	|duxr |dk(  xs |	|k7  }t        j                  ||fdddœ|¤Ž| _         ||fi |¤Ž| _	         |
d¬	«      | _
        t        j                  ||fd
|rdn||	|	|ddœ|¤Ž| _         ||fi |¤Ž| _        |� |«       nt        j                  «       | _         |
d¬	«      | _        t!        |f|||dœ|¤Ž| _        t        j                  ||fdddœ|¤Ž| _         ||fi |¤Ž| _        t)        ||fi |¤Ž| _         |
d¬	«      | _        || _        || _        || _        || _        y)r?   r@   r0   Nr)   r   F)r$   rD   TrE   rC   )r$   r%   r+   r&   ÚgroupsrD   rG   )rJ   rK   ru   ÚmathÚfloorrL   rM   rN   rO   rP   rR   rT   rU   rQ   r<   rW   r   rS   Úconv3Úbn3r   rV   Úact3r3   r%   r&   r=   )rX   r1   r2   r%   r3   r4   r5   r6   r&   r7   r8   r9   r:   r;   r<   r=   rA   rB   rY   ÚwidthrZ   r[   r\   r]   s                          €r,   rK   zBottleneck.__init__¡   s®  ø€ ðL ¨Ñ/ˆÜ‰ÑÔä”D—J‘J˜v¨°b©Ñ9Ó:¸[ÑHÓIˆØ Ñ,ˆØ˜TŸ^™^Ñ+ˆ	Ø'Ò3¨8ˆØ Ð%ÒU¨6°Q©;Ò+T¸.ÈHÑ:Tˆä—Y‘Y˜x¨ÐWÀ1È5ÑWÐTVÑWˆŒ
Ù˜lÑ1¨bÑ1ˆŒÙ dÔ+ˆŒ	ä—Y‘YØØð

ð Ù‘1 FØ"Ø#ØØñ

ð ñ

ˆŒ
ñ ˜eÑ* rÑ*ˆŒØ*4Ð*@™*œ,ÄbÇkÁkÃmˆŒÙ dÔ+ˆŒ	Ü˜HÐY¨u¸VÈFÑYÐVXÑYˆŒä—Y‘Y˜u iÐQ¸QÀUÑQÈbÑQˆŒ
Ù˜iÑ.¨2Ñ.ˆŒä˜j¨)Ñ:°rÑ:ˆŒá dÔ+ˆŒ	Ø$ˆŒØˆŒØ ˆŒØ"ˆ�r.   c                 óš   — t        | j                  dd«      �4t        j                  j	                  | j                  j
                  «       yyr_   )ra   r‚   rM   rb   rc   r`   rd   s    r,   re   zBottleneck.zero_init_lastï   rf   r.   rg   c                 ó8  — |}| j                  |«      }| j                  |«      }| j                  |«      }| j                  |«      }| j	                  |«      }| j                  |«      }| j                  |«      }| j                  |«      }| j                  |«      }| j                  |«      }| j                  �| j                  |«      }| j                  �| j                  |«      }| j                  �| j                  |«      }||z  }| j                  |«      }|S ri   )rO   rP   rR   rT   rU   r<   rW   rS   r�   r‚   rV   r=   r3   rƒ   rj   s      r,   rl   zBottleneck.forwardô   sì   € Øˆà�J‰J�q‹MˆØ�H‰H�Q‹KˆØ�I‰I�a‹Lˆà�J‰J�q‹MˆØ�H‰H�Q‹KˆØ�O‰O˜AÓˆØ�I‰I�a‹LˆØ�G‰G�A‹Jˆà�J‰J�q‹MˆØ�H‰H�Q‹Kˆà�7‰7ÐØ—‘˜“
ˆAà�>‰>Ð%Ø—‘˜qÓ!ˆAà�?‰?Ð&Ø—‘ xÓ0ˆHØ	ˆX‰ˆØ�I‰I�a‹Lˆàˆr.   rm   rn   rz   s   @r,   r#   r#   š   sh  ø„ ñð €Ið Ø.2Ø Ø Ø !ØØ,0Ø)+¯©Ø*,¯.©.Ø48Ø26Ø48Ø-1ØØñ%L#àðL#ð ðL#ð ð	L#ð
 ! §¡Ñ+ðL#ð ðL#ð ðL#ð ðL#ð ðL#ð % S™MðL#ð ˜BŸI™I‘ðL#ð ˜RŸY™Y™ðL#ð !  b§i¡i¡Ñ1ðL#ð ˜t B§I¡I™Ñ/ðL#ð !  b§i¡i¡Ñ1ðL#ð    §	¡	Ñ*ð!L#ð& 
õ'L#ó\,ð
˜Ÿ™ð ¨%¯,©,÷ r.   r#   Úin_channelsÚout_channelsr7   r9   c	           
      óð   — ||dœ}	|xs t         j                  }|dk(  r|dk(  rdn|}|dkD  r|xs |nd}t        |||«      }
t        j                  t        j                  | ||f||
|ddœ|	¤Ž ||fi |	¤ŽgŽ S )Nr@   r   F)r%   r+   r&   rD   )rM   rt   r-   Ú
SequentialrN   )r‡   rˆ   r$   r%   r&   r7   r9   rA   rB   rY   Úps              r,   Údownsample_convrŒ     s¦   € ð  UÑ	+€BØÒ-œrŸ~™~€JØ ’{ x°1¢}‘!¸+€KØ5@À1²_�nÒ0©È!€NÜ�K ¨Ó8€Aä�=‰=Ü
�	‰	ØØØð		
ð ØØ#Øñ		
ð ñ		
ñ 	�<Ñ& 2Ñ&ðð ð r.   c	           
      óH  — ||dœ}	|xs t         j                  }|dk(  r|nd}
|dk(  r|dk(  rt        j                  «       }n,|
dk(  r|dkD  rt        nt         j                  } |d|
dd¬«      }t        j
                  |t        j                  | |dfddddœ|	¤Ž ||fi |	¤ŽgŽ S )	Nr@   r   r)   TF)Ú	ceil_modeÚcount_include_padr   ©r%   r+   rD   )rM   rt   rQ   r   Ú	AvgPool2drŠ   rN   )r‡   rˆ   r$   r%   r&   r7   r9   rA   rB   rY   Ú
avg_strideÚpoolÚavg_pool_fns                r,   Údownsample_avgr•   2  s­   € ð  UÑ	+€BØÒ-œrŸ~™~€JØ# qš=‘¨a€JØ�‚{�x 1’}Ü�{‰{‹}‰à'1°Q¢¸8Àaº<•mÌRÏ\É\ˆÙ˜1˜j°DÈEÔRˆä�=‰=ØÜ
�	‰	�+˜|¨QÐV°qÀ!È%ÑVÐSUÑVÙ�<Ñ& 2Ñ&ðð ð r.   Ú	drop_probc           	      ód   — dd| rt        t        | dd¬«      nd| rt        t        | dd¬«      gS dgS )zÅCreate DropBlock layer instances for each stage.

    Args:
        drop_prob: Drop probability for DropBlock.

    Returns:
        List of DropBlock partial instances or None for each stage.
    Né   ç      Ð?)r–   Ú
block_sizeÚgamma_scalerC   ç      ð?)r   r   )r–   s    r,   Údrop_blocksr�   M  sK   € ð 	ˆdÙU^Œ” y¸QÈDÕQÐdhÙU^Œ” y¸QÈDÔQðjð jð eiðjð jr.   Ú	block_fns.rH   Úblock_repeatsr1   r6   Úoutput_strideÚdown_kernel_sizeÚavg_downÚdrop_block_rateÚdrop_path_ratec                 ó.  — |
|dœ}g }g }t        |«      }d}d}dx}}t        t        | ||t        |«      «      «      D �]R  \  }\  }}}}d|dz   › �}|dk(  rdnd}||k\  r||z  }d}n||z  }d}|dk7  s|||j                  z  k7  rGt        d|||j                  z  |||||j                  d«      d	œ|¤Ž}|rt        di |¤Žn
t        di |¤Ž}t        d|||d
œ|¤Ž}g }t        |«      D ]e  } | dk(  r|nd}| dk(  r|nd}|	|z  |dz
  z  }!|j                   |||||f||!dkD  rt        |!«      nddœ|¤|¤Ž«       |}||j                  z  }|dz  }Œg |j                  |t        j                  |Ž f«       |j                  t        |||¬«      «       �ŒU ||fS )a  Create ResNet stages with specified block configurations.

    Args:
        block_fns: Block class to use for each stage.
        channels: Number of channels for each stage.
        block_repeats: Number of blocks to repeat for each stage.
        inplanes: Number of input channels.
        reduce_first: Reduction factor for first convolution in each stage.
        output_stride: Target output stride of network.
        down_kernel_size: Kernel size for downsample layers.
        avg_down: Use average pooling for downsample.
        drop_block_rate: DropBlock drop rate.
        drop_path_rate: Drop path rate for stochastic depth.
        **kwargs: Additional arguments passed to block constructors.

    Returns:
        Tuple of stage modules list and feature info list.
    r@   r   r|   r   Úlayerr)   Nr9   )r‡   rˆ   r$   r%   r&   r7   r9   )r6   r&   r<   ç        )r7   r=   ©Únum_chsÚ	reductionÚmoduler*   )ÚsumÚ	enumerateÚzipr�   rL   ÚdictÚgetr•   rŒ   ÚrangeÚappendr   rM   rŠ   )"rž   rH   rŸ   r1   r6   r    r¡   r¢   r£   r¤   rA   rB   ÚkwargsrY   ÚstagesÚfeature_infoÚnet_num_blocksÚnet_block_idxÚ
net_strider&   Úprev_dilationÚ	stage_idxÚblock_fnr2   Ú
num_blocksÚdbÚ
stage_namer%   r3   Údown_kwargsÚblock_kwargsÚblocksÚ	block_idxÚ	block_dprs"                                     r,   Úmake_blocksrÄ   \  sI  € ðB  UÑ	+€BØ€FØ€LÜ˜Ó'€NØ€MØ€JØ Ð €Hˆ}Ü9BÄ3ÀyÐRZÐ\iÔkvð  xGó  lHó  DIó  :Jó ,]Ñ5ˆ	Ñ5�H˜f j°"Ø˜Y¨™]˜OÐ,ˆ
Ø 1’n‘¨!ˆØ˜Ò&Ø˜ÑˆHØ‰Fà˜&Ñ ˆJàˆ
Ø�QŠ;˜( f¨x×/AÑ/AÑ&AÒAÜð 	Ø$Ø# h×&8Ñ&8Ñ8Ø,ØØ!Ø,Ø!Ÿ:™: lÓ3ñ	ð ñ	ˆKñ ;CœÑ6¨+Ò6ÌÑHfÐZeÑHfˆJäÐb¨ÀÐUWÑbÐ[aÑbˆØˆÜ˜zÓ*ò 	ˆIØ'0°A¢~™¸4ˆJØ(¨Aš~‘V°1ˆFØ&¨Ñ6¸.È1Ñ:LÑMˆIØ�M‰M™(ØØØØð		ð
  -Ø1:¸R²œ( 9Ô-ÀTñ	ð ð	ð ñ	ô 	ð %ˆMØ × 2Ñ 2Ñ2ˆHØ˜QÑ‰Mð!	ð$ 	�‰�z¤2§=¡=°&Ð#9Ð:Ô;Ø×ÑœD¨¸ZÐPZÔ[Ö\ðY,]ð\ �<ÐÐr.   c            2       ó  ‡ — e Zd ZdZddddddddd	ddd	d
ej
                  ej                  ddddddddfdeee	f   de
e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dee
edf      deded eeej"                        d!ed"ed#ed$ed%eeeef      f.ˆ fd&„Zej.                  j0                  d>d$ed'dfd(„«       Zej.                  j0                  d?d)ed'eeef   fd*„«       Zej.                  j0                  d>d+ed'dfd,„«       Zej.                  j0                  d?d-ed'eeej"                  f   fd.„«       Zd@deded'dfd/„Z	 	 	 	 	 dAd0ej<                  d1eeeee   f      d2ed3ed4ed5ed'eeej<                     e
ej<                  eej<                     f   f   fd6„Z 	 	 	 dBd1eeee   f   d7ed8ed'ee   fd9„Z!d0ej<                  d'ej<                  fd:„Z"d?d0ej<                  d;ed'ej<                  fd<„Z#d0ej<                  d'ej<                  fd=„Z$ˆ xZ%S )Cr!   aª  ResNet / ResNeXt / SE-ResNeXt / SE-Net

    This class implements all variants of ResNet, ResNeXt, SE-ResNeXt, and SENet that
      * have > 1 stride in the 3x3 conv layer of bottleneck
      * have conv-bn-act ordering

    This ResNet impl supports a number of stem and downsample options based on the v1c, v1d, v1e, and v1s
    variants included in the MXNet Gluon ResNetV1b model. The C and D variants are also discussed in the
    'Bag of Tricks' paper: https://arxiv.org/pdf/1812.01187. The B variant is equivalent to torchvision default.

    ResNet variants (the same modifications can be used in SE/ResNeXt models as well):
      * normal, b - 7x7 stem, stem_width = 64, same as torchvision ResNet, NVIDIA ResNet 'v1.5', Gluon v1b
      * c - 3 layer deep 3x3 stem, stem_width = 32 (32, 32, 64)
      * d - 3 layer deep 3x3 stem, stem_width = 32 (32, 32, 64), average pool in downsample
      * e - 3 layer deep 3x3 stem, stem_width = 64 (64, 64, 128), average pool in downsample
      * s - 3 layer deep 3x3 stem, stem_width = 64 (64, 64, 128)
      * t - 3 layer deep 3x3 stem, stem width = 32 (24, 48, 64), average pool in downsample
      * tn - 3 layer deep 3x3 stem, stem width = 32 (24, 32, 64), average pool in downsample

    ResNeXt
      * normal - 7x7 stem, stem_width = 64, standard cardinality and base widths
      * same c,d, e, s variants as ResNet can be enabled

    SE-ResNeXt
      * normal - 7x7 stem, stem_width = 64
      * same c, d, e, s variants as ResNet can be enabled

    SENet-154 - 3 layer deep 3x3 stem (same as v1c-v1s), stem_width = 64, cardinality=64,
        reduction by 2 on width of first bottleneck convolution, 3x3 downsample convs after first block
    éè  rC   é    Úavgr   r0   Ú F)r0   é€   é   i   Nr§   TÚblockÚlayers.Únum_classesÚin_chansr    Úglobal_poolr4   r5   Ú
stem_widthÚ	stem_typeÚreplace_stem_poolÚblock_reduce_firstr¡   r¢   rH   r8   r9   r;   Ú	drop_rater¤   r£   re   Ú
block_argsc                 ó8  •— t         ‰"| �  «        ||dœ}|xs
 t        «       }|dv sJ ‚|| _        || _        || _        d| _        t        |«      }t        |«      }d|
v }|r|	dz  nd}|r­|	|	f}d|
v r
d|	d	z  z  |	f}t        j                  t        j                  ||d
   dfddddœ|¤Ž ||d
   fi |¤Ž |d¬«      t        j                  |d
   |d   dfddddœ|¤Ž ||d   fi |¤Ž |d¬«      t        j                  |d   |dfddddœ|¤ŽgŽ | _        n!t        j                  ||fdddddœ|¤Ž| _         ||fi |¤Ž| _         |d¬«      | _        t        |dd¬«      g| _        |rct        j                  t!        dt        j                  ||df|rdnddddœ|¤Ž|�t#        |f|ddœ|¤Žnd ||fi |¤Ž |d¬«      g«      Ž | _        n€|�at'        |t        j(                  «      r |d«      | _        nVt        j                  t        j*                  ddd¬«       |d|ddœ|¤ŽgŽ | _        nt        j*                  ddd¬«      | _         t-        t/        |«      «      |«      }t1        ||||f|||||||||||dœ|¤|¤Ž\  }} |D ]  }! | j2                  |!Ž  Œ | j                  j5                  | «       |d   |d   j6                  z  x| _        | _        t=        | j8                  | j                  fd|i|¤Ž\  | _        | _         | jC                  |¬«       y)a	  
        Args:
            block (nn.Module): class for the residual block. Options are BasicBlock, Bottleneck.
            layers (List[int]) : number of layers in each block
            num_classes (int): number of classification classes (default 1000)
            in_chans (int): number of input (color) channels. (default 3)
            output_stride (int): output stride of the network, 32, 16, or 8. (default 32)
            global_pool (str): Global pooling type. One of 'avg', 'max', 'avgmax', 'catavgmax' (default 'avg')
            cardinality (int): number of convolution groups for 3x3 conv in Bottleneck. (default 1)
            base_width (int): bottleneck channels factor. `planes * base_width / 64 * cardinality` (default 64)
            stem_width (int): number of channels in stem convolutions (default 64)
            stem_type (str): The type of stem (default ''):
                * '', default - a single 7x7 conv with a width of stem_width
                * 'deep' - three 3x3 convolution layers of widths stem_width, stem_width, stem_width * 2
                * 'deep_tiered' - three 3x3 conv layers of widths stem_width//4 * 3, stem_width, stem_width * 2
            replace_stem_pool (bool): replace stem max-pooling layer with a 3x3 stride-2 convolution
            block_reduce_first (int): Reduction factor for first convolution output width of residual blocks,
                1 for all archs except senets, where 2 (default 1)
            down_kernel_size (int): kernel size of residual block downsample path,
                1x1 for most, 3x3 for senets (default: 1)
            avg_down (bool): use avg pooling for projection skip connection between stages/downsample (default False)
            act_layer (str, nn.Module): activation layer
            norm_layer (str, nn.Module): normalization layer
            aa_layer (nn.Module): anti-aliasing layer
            drop_rate (float): Dropout probability before classifier, for training (default 0.)
            drop_path_rate (float): Stochastic depth drop-path rate (default 0.)
            drop_block_rate (float): Drop block rate (default 0.)
            zero_init_last (bool): zero-init the last weight in residual path (usually last BN affine weight)
            block_args (dict): Extra kwargs to pass through to block module
        r@   )é   é   rÇ   FÚdeepr)   r0   ÚtieredrC   r|   r   r   r�   TrE   é   )r$   r%   r+   rD   rR   r¨   N)rH   r%   )r$   r%   r+   )r4   r5   r    r6   r¢   r¡   r8   r9   r;   r£   r¤   éÿÿÿÿÚ	pool_type)re   r*   )"rJ   rK   r¯   rÎ   rÏ   rÕ   Úgrad_checkpointingr   r   rM   rŠ   rN   rO   rP   rR   rµ   Úfilterr   ÚmaxpoolÚ
issubclassr‘   Ú	MaxPool2dr   ÚlenrÄ   Ú
add_moduleÚextendrL   Únum_featuresÚhead_hidden_sizer   rÐ   ÚfcÚinit_weights)#rX   rÌ   rÍ   rÎ   rÏ   r    rÐ   r4   r5   rÑ   rÒ   rÓ   rÔ   r¡   r¢   rH   r8   r9   r;   rÕ   r¤   r£   re   rÖ   rA   rB   rY   Ú	deep_stemr1   Ústem_chsrž   Ústage_modulesÚstage_feature_infoÚstager]   s#                                     €r,   rK   zResNet.__init__Õ  s§  ø€ ôt 	‰ÑÔØ¨Ñ/ˆØÒ)¤4£6ˆ
Ø Ñ+Ð+Ð+Ø&ˆÔØ ˆŒØ"ˆŒØ"'ˆÔä! )Ó,ˆ	Ü# JÓ/ˆ
ð ˜iÐ'ˆ	Ù%.�: ’>°BˆÙØ" JÐ/ˆHØ˜9Ñ$Ø ¨q¡Ñ1°:Ð>�ÜŸ™Ü—	‘	˜( H¨Q¡K°ÐZ¸1ÀaÈeÑZÐWYÑZÙ˜8 A™;Ñ-¨"Ñ-Ù $Ô'Ü—	‘	˜( 1™+ x°¡{°AÐ]¸aÈÐQVÑ]ÐZ\Ñ]Ù˜8 A™;Ñ-¨"Ñ-Ù $Ô'Ü—	‘	˜( 1™+ x°ÐZ¸1ÀaÈeÑZÐWYÑZð)\ð ]ˆD�Jô Ÿ™ 8¨XÐlÀ1ÈQÐXYÐ`eÑlÐikÑlˆDŒJÙ˜hÑ-¨"Ñ-ˆŒÙ dÔ+ˆŒ	Ü!¨(¸aÈÔOÐPˆÔñ ÜŸ=™=¬&°Ü—	‘	˜( H¨aÐj¹X¹È1ÐVWÐ^cÑjÐgiÑjØJRÐJ^”	˜(ÐF¨X¸aÑFÀ2ÒFÐdhÙ˜8Ñ* rÑ*Ù $Ô'ð	8ó +ð ˆD�Lð Ð#Ü˜h¬¯©Ô5Ù#+¨A£;�D•Lä#%§=¡=ÜŸ™°¸1ÀaÔHÙ ÐC¨(¸1ÑCÀÑCð3Eð $F�D•Lô  "Ÿ|™|¸À!ÈQÔO�”ð -”Iœc (›mÓ,¨UÓ3ˆ	Ü,7ØØØØð	-
ð
 $Ø!Ø'Ø+ØØ-ØØ!ØØ+Ø)ñ-
ð  ð!-
ð" ñ#-
Ñ)ˆÐ)ð& #ò 	$ˆEØˆD�O‰O˜UÒ#ð	$à×Ñ× Ñ Ð!3Ô4ð 5=¸R±LÀ9ÈRÁ=×CZÑCZÑ4ZÐZˆÔ˜DÔ1Ü$5°d×6GÑ6GÈ×IYÑIYÑ$wÐepÐ$wÐtvÑ$wÑ!ˆÔ˜$œ'à×Ñ¨ÐÕ8r.   r'   c                 ó,  — | j                  «       D ]L  \  }}t        |t        j                  «      sŒ!t        j                  j                  |j                  dd¬«       ŒN |r3| j                  «       D ]  }t        |d«      sŒ|j                  «        Œ! yy)z‚Initialize model weights.

        Args:
            zero_init_last: Zero-initialize the last BN in each residual branch.
        Úfan_outÚrelu)ÚmodeÚnonlinearityre   N)
Únamed_modulesÚ
isinstancerM   rN   rb   Úkaiming_normal_r`   ÚmodulesÚhasattrre   )rX   re   ÚnÚms       r,   rê   zResNet.init_weightsb  sƒ   € ð ×&Ñ&Ó(ò 	W‰DˆAˆqÜ˜!œRŸY™YÕ'Ü—‘×'Ñ'¨¯©°yÈvÐ'ÕVð	Wñ Ø—\‘\“^ò '�Ü˜1Ð.Õ/Ø×$Ñ$Õ&ñ'ð r.   Úcoarsec                 ó(   — t        d|rdnd¬«      }|S )zâCreate regex patterns for parameter grouping.

        Args:
            coarse: Use coarse (stage-level) or fine (block-level) grouping.

        Returns:
            Dictionary mapping group names to regex patterns.
        z^conv1|bn1|maxpoolz^layer(\d+)z^layer(\d+)\.(\d+))ÚstemrÁ   )r¯   )rX   rü   Úmatchers      r,   Úgroup_matcherzResNet.group_matcherq  s   € ô Ð1ÉF¹.ÐXmÔnˆØˆr.   rI   c                 ó   — || _         y)zEnable or disable gradient checkpointing.

        Args:
            enable: Whether to enable gradient checkpointing.
        N)rß   )rX   rI   s     r,   Úset_grad_checkpointingzResNet.set_grad_checkpointing~  s   € ð #)ˆÕr.   Ú	name_onlyc                 ó"   — |rdS | j                   S )z³Get the classifier module.

        Args:
            name_only: Return classifier module name instead of module.

        Returns:
            Classifier module or name.
        ré   )ré   )rX   r  s     r,   Úget_classifierzResNet.get_classifier‡  s   € ñ !ˆtÐ- d§g¡gÐ-r.   c                 óp   — || _         t        | j                  | j                   |¬«      \  | _        | _        y)zŸReset the classifier head.

        Args:
            num_classes: Number of classes for new classifier.
            global_pool: Global pooling type.
        )rÞ   N)rÎ   r   rç   rÐ   ré   )rX   rÎ   rÐ   s      r,   Úreset_classifierzResNet.reset_classifier“  s1   € ð 'ˆÔÜ$5°d×6GÑ6GÈ×IYÑIYÐepÔ$qÑ!ˆÔ˜$�'r.   rg   ÚindicesÚnormÚ
stop_earlyÚ
output_fmtÚintermediates_onlyc                 óz  — |dv sJ d«       ‚g }t        d|«      \  }}	d}
| j                  |«      }| j                  |«      }| j                  |«      }|
|v r|j	                  |«       | j                  |«      }d}|r|d|	 }|D ]/  }|
dz  }
 t        | |«      |«      }|
|v sŒ|j	                  |«       Œ1 |r|S ||fS )aJ  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:
            Features and list of intermediate features or just intermediate features.
        )ÚNCHWzOutput shape must be NCHW.r˜   r   ©Úlayer1Úlayer2Úlayer3Úlayer4Nr   )r   rO   rP   rR   r²   rá   ra   )rX   rg   r  r	  r
  r  r  ÚintermediatesÚtake_indicesÚ	max_indexÚfeat_idxÚlayer_namesrú   s                r,   Úforward_intermediateszResNet.forward_intermediates�  sê   € ð, ˜YÑ&ÐDÐ(DÓDÐ&ØˆÜ"6°q¸'Ó"BÑˆ�ið ˆØ�J‰J�q‹MˆØ�H‰H�Q‹KˆØ�I‰I�a‹LˆØ�|Ñ#Ø× Ñ  Ô#Ø�L‰L˜‹Oˆà>ˆÙØ% j yÐ1ˆKØò 	(ˆAØ˜‰MˆHØ ”˜˜aÓ  Ó#ˆAØ˜<Ò'Ø×$Ñ$ QÕ'ð		(ñ Ø Ð à�-ÐÐr.   Ú
prune_normÚ
prune_headc                 ó¦   — t        d|«      \  }}d}||d }|D ]!  }t        | |t        j                  «       «       Œ# |r| j	                  dd«       |S )aF  Prune layers not required for specified intermediates.

        Args:
            indices: Indices of intermediate layers to keep.
            prune_norm: Whether to prune normalization layers.
            prune_head: Whether to prune the classifier head.

        Returns:
            List of indices that were kept.
        r˜   r  Nr   rÉ   )r   ÚsetattrrM   rQ   r  )rX   r  r  r  r  r  r  rú   s           r,   Úprune_intermediate_layersz ResNet.prune_intermediate_layersÎ  s`   € ô  #7°q¸'Ó"BÑˆ�iØ>ˆØ! ) *Ð-ˆØò 	,ˆAÜ�D˜!œRŸ[™[›]Õ+ð	,áØ×!Ñ! ! RÔ(ØÐr.   c                 óâ  — | j                  |«      }| j                  |«      }| j                  |«      }| j                  |«      }| j                  rZt
        j                  j                  «       s<t        | j                  | j                  | j                  | j                  g|d¬«      }|S | j                  |«      }| j                  |«      }| j                  |«      }| j                  |«      }|S )z/Forward pass through feature extraction layers.T)Úflatten)rO   rP   rR   rá   rß   rw   ÚjitÚis_scriptingr   r  r  r  r  ©rX   rg   s     r,   Úforward_featureszResNet.forward_featuresç  s¹   € à�J‰J�q‹MˆØ�H‰H�Q‹KˆØ�I‰I�a‹LˆØ�L‰L˜‹Oˆà×"Ò"¬5¯9©9×+AÑ+AÔ+CÜ §¡¨T¯[©[¸$¿+¹+ÀtÇ{Á{ÐSÐUVÐ`dÔeˆAð ˆð	 —‘˜A“ˆAØ—‘˜A“ˆAØ—‘˜A“ˆAØ—‘˜A“ˆAØˆr.   Ú
pre_logitsc                 óÐ   — | j                  |«      }| j                  r5t        j                  |t	        | j                  «      | j
                  ¬«      }|r|S | j                  |«      S )zÐForward pass through classifier head.

        Args:
            x: Feature tensor.
            pre_logits: Return features before final classifier layer.

        Returns:
            Output tensor.
        )r‹   Útraining)rÐ   rÕ   ÚFÚdropoutÚfloatr'  ré   )rX   rg   r%  s      r,   Úforward_headzResNet.forward_head÷  sO   € ð ×Ñ˜QÓˆØ�>Š>Ü—	‘	˜!œu T§^¡^Ó4¸t¿}¹}ÔMˆAÙˆqÐ. D§G¡G¨A£JÐ.r.   c                 óJ   — | j                  |«      }| j                  |«      }|S )zForward pass.)r$  r+  r#  s     r,   rl   zResNet.forward  s'   € à×!Ñ! !Ó$ˆØ×Ñ˜aÓ ˆØˆr.   )T©F)rÈ   )NFFr  F)r   FT)&ro   rp   rq   rr   rM   rs   rt   r
   r"   r#   r   ru   ÚstrÚboolr   r   r	   rv   r*  r   r   rK   rw   r!  Úignorerê   r   r  r  r  rx   r   r  r  r$  r+  rl   ry   rz   s   @r,   r!   r!   µ  s°  ø„ ñðF  $ØØ!#Ø$Ø Ø Ø ØØ&+Ø&'Ø$%Ø"Ø2EØ#%§7¡7Ø$&§N¡NØ26Ø"Ø$&Ø%'Ø#'Ø37ØØñ5K9à˜ ZÐ/Ñ0ðK9ð ˜#˜s˜(‘OðK9ð ð	K9ð
 ðK9ð ðK9ð ðK9ð ðK9ð ðK9ð ðK9ð ðK9ð  $ðK9ð !$ðK9ð "ðK9ð ðK9ð  ˜u S¨# X™Ñ/ð!K9ð" !ð#K9ð$ "ð%K9ð& ˜t B§I¡I™Ñ/ð'K9ð( ð)K9ð* "ð+K9ð, #ð-K9ð. !ð/K9ð0 !  c¨3 h¡Ñ0õ1K9ðZ ‡Y�Y×Ññ'¨4ð '¸4ò 'ó ð'ð ‡Y�Y×Ññ
 Dð 
°T¸#¸s¸(±^ò 
ó ð
ð ‡Y�Y×Ññ)¨Tð )¸Tò )ó ð)ð ‡Y�Y×Ññ	.¨ð 	.¸¸sÀBÇIÁI¸~Ñ9Nò 	.ó ð	.ñr¨Cð r¸cð rÈdó rð 8<ØØ$Ø$Ø',ñ/ à�|‰|ð/ ð ˜e C¨¨c© NÑ3Ñ4ð/ ð ð	/ ð
 ð/ ð ð/ ð !%ð/ ð 
ˆt�E—L‘LÑ! 5¨¯©°t¸E¿L¹LÑ7IÐ)IÑ#JÐJÑ	Kó/ ðf ./Ø$Ø#ñ	à˜3  S¡	˜>Ñ*ðð ðð ð	ð
 
ˆc‰óð2 %§,¡,ð °5·<±<ó ñ /˜eŸl™lð /¸ð /ÈÏÉó /ð˜Ÿ™ð ¨%¯,©,÷ r.   r!   ÚvariantÚ
pretrainedc                 ó&   — t        t        | |fi |¤ŽS )zÓCreate a ResNet model.

    Args:
        variant: Model variant name.
        pretrained: Load pretrained weights.
        **kwargs: Additional model arguments.

    Returns:
        ResNet model instance.
    )r   r!   )r1  r2  r³   s      r,   Ú_create_resnetr4    s   € ô  ¤¨°ÑF¸vÑFÐFr.   Úurlc                 ó2   — | dddddt         t        dddd	œ|¥S )
z1Create a default configuration for ResNet models.rÆ   ©rC   éà   r8  )rÜ   rÜ   g      ì?ÚbilinearrO   ré   z
apache-2.0)r5  rÎ   Ú
input_sizeÚ	pool_sizeÚcrop_pctÚinterpolationÚmeanÚstdÚ
first_convÚ
classifierÚlicenser   ©r5  r³   s     r,   Ú_cfgrD    s5   € ð Ø¨=ÀvØ¨JÜ%Ô.BØ¨TØñð ðð r.   c           	      ó4   — t        dd| it        ddifi |¤Ž¤ŽS )z2Create a configuration with bicubic interpolation.r5  r=  Úbicubicr*   ©rD  r¯   rC  s     r,   Ú_tcfgrH  (  s$   € äÑH�CÐHœ4 °)Ð <ÑGÀÑGÑHÐHr.   c                 ó:   — t        dd| it        dddddœfi |¤Ž¤ŽS )z4Create a configuration for models trained with timm.r5  rF  ©rC   é   rK  çffffffî?ú3https://github.com/huggingface/pytorch-image-models)r=  Útest_input_sizeÚtest_crop_pctÚ
origin_urlr*   rG  rC  s     r,   Ú_ttcfgrQ  -  s<   € äñ �Cð œ4Ø"°}ÐW[ØKñ!ñ ð 
ññ ð r.   c                 ó>   — t        d	d| it        dddddddœfi |¤Ž¤ŽS )
z,Create a configuration for ResNet-RS models.r5  rF  rL  rJ  rœ   rM  úarXiv:2110.00476)r=  r<  rN  rO  rP  Ú	paper_idsr*   rG  rC  s     r,   Ú_rcfgrU  5  sA   € äñ �Cð œ4Ø"°ÈÐilØKÐZlñ!ñ ð 
ññ ð r.   c                 óB   — t        d
d| it        ddddddddd	œfi |¤Ž¤ŽS )z?Create a configuration for ResNet-RS models with 160x160 input.r5  rF  ©rC   é    rX  ©r˜   r˜   rL  r7  rM  rS  )r=  r:  r;  r<  rN  rO  rP  rT  r*   rG  rC  s     r,   Ú_r3cfgrZ  =  sE   € äñ �Cð œ4Ø"°-ÈfØ¨]ÈTØKÐZlñ!ñ ð 
ñ	ñ ð r.   c           	      ó6   — t        dd| it        dddœfi |¤Ž¤ŽS )z3Create a configuration for Gluon pretrained models.r5  rF  z1https://cv.gluon.ai/model_zoo/classification.html)r=  rP  r*   rG  rC  s     r,   Ú_gcfgr\  F  s7   € äñ �Cð œ4Ø"ØIñ!ñ ð 
ññ ð r.   zresnet10t.c3_in1kztimm/zrhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet10t_176_c3-f3215ab1.pth)rC   é°   r]  )é   r^  rL  r7  zconv1.0)Ú	hf_hub_idr5  r:  r;  rO  rN  r@  zresnet14t.c3_in1kzrhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet14t_176_c3-c4ed2c37.pthzresnet18.a1_in1kzqhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet18_a1_0-d63eafa0.pth)r_  r5  zresnet18.a2_in1kzqhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet18_a2_0-b61bd467.pthzresnet18.a3_in1kzqhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet18_a3_0-40c531c8.pthzresnet18d.ra2_in1kzkhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet18d_ra2-48a79e06.pth)r_  r5  r@  zresnet18d.ra4_e3600_r224_in1k)ç      à?r`  r`  gÍÌÌÌÌÌì?)r_  r>  r?  r<  r@  zresnet34.a1_in1kzqhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet34_a1_0-46f8f793.pthzresnet34.a2_in1kzqhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet34_a2_0-82d47d71.pthzresnet34.a3_in1kzqhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet34_a3_0-a20cabb6.pth)r_  r5  r<  zresnet34.bt_in1kzfhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet34-43635321.pthzresnet34.ra4_e3600_r224_in1k)r_  r>  r?  r<  zresnet34d.ra2_in1kzkhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet34d_ra2-f8dcfcaf.pthzresnet26.bt_in1kzfhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet26-9aa10e23.pthzresnet26d.bt_in1kzghttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet26d-69e92c46.pthzresnet26t.ra2_in1kzthttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-attn-weights/resnet26t_256_ra2-6f6fa748.pth)rC   rË   rË   )rØ   rØ   g®Gázî?)rC   é@  ra  rœ   )r_  r5  r@  r:  r;  r<  rN  rO  zresnet50.a1_in1kzohttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_a1_0-14fe96d1.pthzresnet50.a1h_in1kzuhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_a1h2_176-001a1197.pth)r_  r5  r:  r;  r<  rN  rO  zresnet50.a2_in1kzqhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_a2_0-a2746f79.pthzresnet50.a3_in1kzqhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_a3_0-59cae1ef.pthzresnet50.b1k_in1kzphttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_b1k-532a802a.pthzresnet50.b2k_in1kzphttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_b2k-1ba180c1.pthzresnet50.c1_in1kzohttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_c1-5ba5e060.pthzresnet50.c2_in1kzohttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_c2-d01e05b2.pthzresnet50.d_in1kznhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_d-f39db8af.pthzresnet50.ram_in1kzlhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/resnet50_ram-a26f946b.pthzresnet50.am_in1kzkhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/resnet50_am-6c502b37.pthzresnet50.ra_in1kzkhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/resnet50_ra-85ebb6e5.pthzresnet50.bt_in1kzkhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/rw_resnet50-86acaeed.pthzresnet50d.ra2_in1kzkhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet50d_ra2-464e36ba.pthzresnet50d.ra4_e3600_r224_in1krJ  )r_  r>  r?  r<  rN  rO  r@  zresnet50d.a1_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50d_a1_0-e20cff14.pthzresnet50d.a2_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50d_a2_0-a3adc64d.pthzresnet50d.a3_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50d_a3_0-403fdfad.pthzresnet50t.untrained)r@  zresnet101.a1h_in1kzohttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet101_a1h-36d3f2aa.pthzresnet101.a1_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet101_a1_0-cdcb52a9.pthzresnet101.a2_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet101_a2_0-6edb36c7.pthzresnet101.a3_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet101_a3_0-1db14157.pthzresnet101d.ra2_in1kzlhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet101d_ra2-2803ffab.pth)r_  r5  r@  r:  r;  r<  rO  rN  zresnet152.a1h_in1kzohttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet152_a1h-dc400468.pthzresnet152.a1_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet152_a1_0-2eee8a7a.pthzresnet152.a2_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet152_a2_0-b4c6978f.pthzresnet152.a3_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet152_a3_0-134d4688.pthzresnet152d.ra2_in1kzlhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet152d_ra2-5cac0439.pthzresnet200.untrainedzresnet200d.ra2_in1kzlhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnet200d_ra2-bdba9bf9.pthzwide_resnet50_2.racm_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/wide_resnet50_racm-8234f177.pthzresnet18.tv_in1kz9https://download.pytorch.org/models/resnet18-f37072fd.pthzbsd-3-clausez!https://github.com/pytorch/vision)r_  r5  rB  rP  zresnet34.tv_in1kz9https://download.pytorch.org/models/resnet34-b627a593.pthzresnet50.tv_in1kz9https://download.pytorch.org/models/resnet50-0676ba61.pthzresnet50.tv2_in1kz9https://download.pytorch.org/models/resnet50-11ad3fa6.pthgáz®Gáî?)r_  r5  r:  r;  rN  rO  rB  rP  zresnet101.tv_in1kz:https://download.pytorch.org/models/resnet101-63fe2227.pthzresnet101.tv2_in1kz:https://download.pytorch.org/models/resnet101-cd907fc2.pthzresnet152.tv_in1kz:https://download.pytorch.org/models/resnet152-394f9c45.pthzresnet152.tv2_in1kz:https://download.pytorch.org/models/resnet152-f82ba261.pthzwide_resnet50_2.tv_in1kz@https://download.pytorch.org/models/wide_resnet50_2-95faca4d.pthzwide_resnet50_2.tv2_in1kz@https://download.pytorch.org/models/wide_resnet50_2-9ba9bcbe.pthzwide_resnet101_2.tv_in1kzAhttps://download.pytorch.org/models/wide_resnet101_2-32ee1156.pthzwide_resnet101_2.tv2_in1kzAhttps://download.pytorch.org/models/wide_resnet101_2-d733dc28.pthzresnet50_gn.a1h_in1kzrhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/resnet50_gn_a1h2-8fe6c4d0.pthzresnext50_32x4d.a1h_in1kzuhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/resnext50_32x4d_a1h-0146ab0a.pthzresnext50_32x4d.a1_in1kzxhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnext50_32x4d_a1_0-b5a91a1d.pthzresnext50_32x4d.a2_in1kzxhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnext50_32x4d_a2_0-efc76add.pthzresnext50_32x4d.a3_in1kzxhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/resnext50_32x4d_a3_0-3e450271.pthzresnext50_32x4d.ra_in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/resnext50_32x4d_ra-d733960d.pthzresnext50d_32x4d.bt_in1kznhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnext50d_32x4d-103e99f8.pthzresnext101_32x4d.untrainedzresnext101_64x4d.c1_in1kzthttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tpu-weights/resnext101_64x4d_c-0d0e0cc0.pthzresnext50_32x4d.tv_in1kz@https://download.pytorch.org/models/resnext50_32x4d-7cdf4587.pthzresnext101_32x8d.tv_in1kzAhttps://download.pytorch.org/models/resnext101_32x8d-8ba56ff5.pthzresnext101_64x4d.tv_in1kzAhttps://download.pytorch.org/models/resnext101_64x4d-173b62eb.pthzresnext50_32x4d.tv2_in1kz@https://download.pytorch.org/models/resnext50_32x4d-1a0047aa.pthzresnext101_32x8d.tv2_in1kzAhttps://download.pytorch.org/models/resnext101_32x8d-110c445d.pthz$resnext101_32x8d.fb_wsl_ig1b_ft_in1kzChttps://download.pytorch.org/models/ig_resnext101_32x8-c38310e5.pthzcc-by-nc-4.0z.https://github.com/facebookresearch/WSL-Imagesz%resnext101_32x16d.fb_wsl_ig1b_ft_in1kzDhttps://download.pytorch.org/models/ig_resnext101_32x16-c6f796b0.pthz%resnext101_32x32d.fb_wsl_ig1b_ft_in1kzDhttps://download.pytorch.org/models/ig_resnext101_32x32-e4b90b00.pthz%resnext101_32x48d.fb_wsl_ig1b_ft_in1kzDhttps://download.pytorch.org/models/ig_resnext101_32x48-3e41cc8a.pthz resnet18.fb_ssl_yfcc100m_ft_in1kzdhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_supervised_resnet18-d92f0530.pthzEhttps://github.com/facebookresearch/semi-supervised-ImageNet1K-modelsz resnet50.fb_ssl_yfcc100m_ft_in1kzdhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_supervised_resnet50-08389792.pthz'resnext50_32x4d.fb_ssl_yfcc100m_ft_in1kzjhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_supervised_resnext50_32x4-ddb3e555.pthz(resnext101_32x4d.fb_ssl_yfcc100m_ft_in1kzkhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_supervised_resnext101_32x4-dc43570a.pthz(resnext101_32x8d.fb_ssl_yfcc100m_ft_in1kzkhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_supervised_resnext101_32x8-2cfe2f8b.pthz)resnext101_32x16d.fb_ssl_yfcc100m_ft_in1kzlhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_supervised_resnext101_32x16-15fffa57.pthzresnet18.fb_swsl_ig1b_ft_in1kzkhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_weakly_supervised_resnet18-118f1556.pthzresnet50.fb_swsl_ig1b_ft_in1kzkhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_weakly_supervised_resnet50-16a12f1b.pthz$resnext50_32x4d.fb_swsl_ig1b_ft_in1kzqhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_weakly_supervised_resnext50_32x4-72679e44.pthz%resnext101_32x4d.fb_swsl_ig1b_ft_in1kzrhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_weakly_supervised_resnext101_32x4-3f87e46b.pthz%resnext101_32x8d.fb_swsl_ig1b_ft_in1kzrhttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_weakly_supervised_resnext101_32x8-b4712904.pthz&resnext101_32x16d.fb_swsl_ig1b_ft_in1kzshttps://dl.fbaipublicfiles.com/semiweaksupervision/model_files/semi_weakly_supervised_resnext101_32x16-f3559a9c.pthzecaresnet26t.ra2_in1kznhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/ecaresnet26t_ra2-46609757.pth)r_  r5  r@  r:  r;  rO  rN  zecaresnetlight.miil_in1kzlhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tresnet/ecaresnetlight-75a9c627.pth)r_  r5  rO  rN  zecaresnet50d.miil_in1kzjhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tresnet/ecaresnet50d-93c81e3b.pth)r_  r5  r@  rO  rN  zecaresnet50d_pruned.miil_in1kzlhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tresnet/ecaresnet50d_p-e4fa23c2.pthzecaresnet50t.ra2_in1kznhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/ecaresnet50t_ra2-f7ac63c4.pthzecaresnet50t.a1_in1kzuhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/ecaresnet50t_a1_0-99bd76a8.pthzecaresnet50t.a2_in1kzuhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/ecaresnet50t_a2_0-b1c7b745.pthzecaresnet50t.a3_in1kzuhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/ecaresnet50t_a3_0-8cc311f1.pthzecaresnet101d.miil_in1kzkhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tresnet/ecaresnet101d-153dad65.pthzecaresnet101d_pruned.miil_in1kzmhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tresnet/ecaresnet101d_p-9e74cb91.pthzecaresnet200d.untrained)r@  r:  r<  r;  zecaresnet269d.ra2_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/ecaresnet269d_320_ra2-7baa55cb.pth)é
   rb  )rC   é`  rc  zecaresnext26t_32x4d.untrainedzecaresnext50t_32x4d.untrainedzseresnet18.untrainedzseresnet34.untrainedzseresnet50.a1_in1kzshttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/seresnet50_a1_0-ffa00869.pthzseresnet50.a2_in1kzshttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/seresnet50_a2_0-850de0d9.pthzseresnet50.a3_in1kzshttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-rsb-weights/seresnet50_a3_0-317ecd56.pthzseresnet50.ra2_in1kzohttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/seresnet50_ra_224-8efdb4bb.pthzseresnet50t.untrainedzseresnet101.untrainedzseresnet152.untrainedzseresnet152d.ra2_in1kznhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/seresnet152d_ra2-04464dd2.pthzseresnet200d.untrained)r@  r:  r;  zseresnet269d.untrainedzseresnext26d_32x4d.bt_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/seresnext26d_32x4d-80fa48a3.pthzseresnext26t_32x4d.bt_in1kzqhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/seresnext26tn_32x4d-569cb627.pthzseresnext50_32x4d.racm_in1kzthttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/seresnext50_32x4d_racm-a304a460.pthzseresnext101_32x4d.untrainedzseresnext101_32x8d.ah_in1kzwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tpu-weights/seresnext101_32x8d_ah-e6bc4c0a.pthzseresnext101d_32x8d.ah_in1kzxhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tpu-weights/seresnext101d_32x8d_ah-191d7b94.pthzresnetaa50d.sw_in12k_ft_in1k)r_  r@  r<  rO  zresnetaa101d.sw_in12k_ft_in1kz*seresnextaa101d_32x8d.sw_in12k_ft_in1k_288)é	   rd  )r_  r<  r:  r;  rN  rO  r@  z&seresnextaa101d_32x8d.sw_in12k_ft_in1k)r_  r@  rO  z*seresnextaa201d_32x8d.sw_in12k_ft_in1k_384rF  )é   re  )rC   é€  rf  )r_  r=  r@  r;  r:  r<  zseresnextaa201d_32x8d.sw_in12ki-.  )	r_  rÎ   r=  r@  r<  r:  r;  rN  rO  zresnetaa50d.sw_in12k)r_  rÎ   r@  r<  rO  zresnetaa50d.d_in12kzresnetaa101d.sw_in12kzseresnextaa101d_32x8d.sw_in12kzresnetblur18.untrainedzresnetblur50.bt_in1kzjhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/resnetblur50-84f4748f.pthzresnetblur50d.untrainedzresnetblur101d.untrainedzresnetaa34d.untrainedzresnetaa50.a1h_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/resnetaa50_a1h-4cf422b3.pthzseresnetaa50d.untrainedzseresnextaa101d_32x8d.ah_in1kzzhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tpu-weights/seresnextaa101d_32x8d_ah-83c8ae12.pthzresnetrs50.tf_in1kzohttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rs-weights/resnetrs50_ema-6b53758b.pthrW  rY  g…ëQ¸í?)r_  r5  r:  r;  r<  rN  r=  r@  zresnetrs101.tf_in1kzuhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rs-weights/resnetrs101_i192_ema-1509bbf6.pth)rC   éÀ   rg  zresnetrs152.tf_in1kzuhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rs-weights/resnetrs152_i256_ema-a9aff7f9.pthzresnetrs200.tf_in1kzohttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tpu-weights/resnetrs200_c-6b698b88.pthzresnetrs270.tf_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rs-weights/resnetrs270_ema-b40e674c.pthzresnetrs350.tf_in1kzuhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rs-weights/resnetrs350_i256_ema-5a1aa8f1.pthzresnetrs420.tf_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rs-weights/resnetrs420_ema-972dee69.pth)rC   é   rh  zresnet18.gluon_in1kzrhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet18_v1b-0757602b.pthzresnet34.gluon_in1kzrhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet34_v1b-c6d82d59.pthzresnet50.gluon_in1kzrhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet50_v1b-0ebe02e2.pthzresnet101.gluon_in1kzshttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet101_v1b-3b017079.pthzresnet152.gluon_in1kzshttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet152_v1b-c1edb0dd.pthzresnet50c.gluon_in1kzrhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet50_v1c-48092f55.pthzresnet101c.gluon_in1kzshttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet101_v1c-1f26822a.pthzresnet152c.gluon_in1kzshttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet152_v1c-a3bb0b98.pthzresnet50d.gluon_in1kzrhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet50_v1d-818a1b1b.pthzresnet101d.gluon_in1kzshttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet101_v1d-0f9c8644.pthzresnet152d.gluon_in1kzshttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet152_v1d-bd354e12.pthzresnet50s.gluon_in1kzrhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet50_v1s-1762acc0.pthzresnet101s.gluon_in1kzshttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet101_v1s-60fe0cc1.pthzresnet152s.gluon_in1kzshttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnet152_v1s-dcc41b81.pthzresnext50_32x4d.gluon_in1kzuhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnext50_32x4d-e6a097c1.pthzresnext101_32x4d.gluon_in1kzvhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnext101_32x4d-b253c8c4.pthzresnext101_64x4d.gluon_in1kzvhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_resnext101_64x4d-f9a8e184.pthzseresnext50_32x4d.gluon_in1kzwhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_seresnext50_32x4d-90cf2d6e.pthzseresnext101_32x4d.gluon_in1kzxhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_seresnext101_32x4d-cf52900d.pthzseresnext101_64x4d.gluon_in1kzxhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_seresnext101_64x4d-f9926f93.pthzsenet154.gluon_in1kznhttps://github.com/rwightman/pytorch-pretrained-gluonresnet/releases/download/v0.1/gluon_senet154-70a1a3c0.pthztest_resnet.r160_in1k)r_  r>  r?  r<  r:  r;  r@  c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-10-T model.
    ©r   r   r   r   rÇ   Údeep_tieredT©rÌ   rÍ   rÑ   rÒ   r¢   Ú	resnet10t©r¯   r"   r4  ©r2  r³   Ú
model_argss      r,   rm  rm  Ä  ó4   € ô œJ¨|ÈÐVcÐnrÔs€JÜ˜+ zÑP´T¸*Ñ5OÈÑ5OÑPÐPr.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-14-T model.
    rj  rÇ   rk  Trl  Ú	resnet14t©r¯   r#   r4  ro  s      r,   rs  rs  Ì  rq  r.   c           	      óR   — t        t        d¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z"Constructs a ResNet-18 model.
    ©r)   r)   r)   r)   ©rÌ   rÍ   Úresnet18rn  ro  s      r,   rx  rx  Ô  ó,   € ô œJ¨|Ô<€JÜ˜* jÑO´D¸Ñ4NÀvÑ4NÑOÐOr.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-18-D model.
    rv  rÇ   rÚ   Trl  Ú	resnet18drn  ro  s      r,   r{  r{  Ü  ó4   € ô œJ¨|ÈÐV\ÐgkÔl€JÜ˜+ zÑP´T¸*Ñ5OÈÑ5OÑPÐPr.   c           	      óR   — t        t        d¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z"Constructs a ResNet-34 model.
    ©rC   r|   r^  rC   rw  Úresnet34rn  ro  s      r,   r  r  ä  ry  r.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-34-D model.
    r~  rÇ   rÚ   Trl  Ú	resnet34drn  ro  s      r,   r�  r�  ì  r|  r.   c           	      óR   — t        t        d¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z"Constructs a ResNet-26 model.
    rv  rw  Úresnet26rt  ro  s      r,   rƒ  rƒ  ô  ry  r.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-26-T model.
    rv  rÇ   rk  Trl  Ú	resnet26trt  ro  s      r,   r…  r…  ü  rq  r.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-26-D model.
    rv  rÇ   rÚ   Trl  Ú	resnet26drt  ro  s      r,   r‡  r‡    r|  r.   c           	      óR   — t        t        d¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z"Constructs a ResNet-50 model.
    r~  rw  Úresnet50rt  ro  s      r,   r‰  r‰    ry  r.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-50-C model.
    r~  rÇ   rÚ   ©rÌ   rÍ   rÑ   rÒ   Ú	resnet50crt  ro  s      r,   rŒ  rŒ    ó1   € ô œJ¨|ÈÐV\Ô]€JÜ˜+ zÑP´T¸*Ñ5OÈÑ5OÑPÐPr.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-50-D model.
    r~  rÇ   rÚ   Trl  Ú	resnet50drt  ro  s      r,   r�  r�    r|  r.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-50-S model.
    r~  r0   rÚ   r‹  Ú	resnet50srt  ro  s      r,   r‘  r‘  $  r�  r.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z$Constructs a ResNet-50-T model.
    r~  rÇ   rk  Trl  Ú	resnet50trt  ro  s      r,   r“  r“  ,  rq  r.   c           	      óR   — t        t        d¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z#Constructs a ResNet-101 model.
    ©rC   r|   é   rC   rw  Ú	resnet101rt  ro  s      r,   r—  r—  4  ó,   € ô œJ¨}Ô=€JÜ˜+ zÑP´T¸*Ñ5OÈÑ5OÑPÐPr.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z%Constructs a ResNet-101-C model.
    r•  rÇ   rÚ   r‹  Ú
resnet101crt  ro  s      r,   rš  rš  <  ó1   € ô œJ¨}ÈÐW]Ô^€JÜ˜,¨
ÑQ´d¸:Ñ6PÈÑ6PÑQÐQr.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z%Constructs a ResNet-101-D model.
    r•  rÇ   rÚ   Trl  Ú
resnet101drt  ro  s      r,   r�  r�  D  ó4   € ô œJ¨}ÈÐW]ÐhlÔm€JÜ˜,¨
ÑQ´d¸:Ñ6PÈÑ6PÑQÐQr.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z%Constructs a ResNet-101-S model.
    r•  r0   rÚ   r‹  Ú
resnet101srt  ro  s      r,   r   r   L  r›  r.   c           	      óR   — t        t        d¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z#Constructs a ResNet-152 model.
    ©rC   rØ   é$   rC   rw  Ú	resnet152rt  ro  s      r,   r¤  r¤  T  r˜  r.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z%Constructs a ResNet-152-C model.
    r¢  rÇ   rÚ   r‹  Ú
resnet152crt  ro  s      r,   r¦  r¦  \  r›  r.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z%Constructs a ResNet-152-D model.
    r¢  rÇ   rÚ   Trl  Ú
resnet152drt  ro  s      r,   r¨  r¨  d  rž  r.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z%Constructs a ResNet-152-S model.
    r¢  r0   rÚ   r‹  Ú
resnet152srt  ro  s      r,   rª  rª  l  r›  r.   c           	      óR   — t        t        d¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z#Constructs a ResNet-200 model.
    ©rC   é   r£  rC   rw  Ú	resnet200rt  ro  s      r,   r®  r®  t  s,   € ô œJ¨~Ô>€JÜ˜+ zÑP´T¸*Ñ5OÈÑ5OÑPÐPr.   c           	      óX   — t        t        dddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z%Constructs a ResNet-200-D model.
    r¬  rÇ   rÚ   Trl  Ú
resnet200drt  ro  s      r,   r°  r°  |  s4   € ô œJ¨~È"ÐX^ÐimÔn€JÜ˜,¨
ÑQ´d¸:Ñ6PÈÑ6PÑQÐQr.   c           	      óT   — t        t        dd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )aO  Constructs a Wide ResNet-50-2 model.
    The model is the same as ResNet except for the bottleneck number of channels
    which is twice larger in every block. The number of channels in outer 1x1
    convolutions is the same, e.g. last block in ResNet-50 has 2048-512-2048
    channels, and in Wide ResNet-50-2 has 2048-1024-2048.
    r~  rÊ   ©rÌ   rÍ   r5   Úwide_resnet50_2rt  ro  s      r,   r³  r³  „  s/   € ô œJ¨|ÈÔL€JÜÐ+¨ZÑV¼4À
Ñ;UÈfÑ;UÑVÐVr.   c           	      óT   — t        t        dd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )zçConstructs a Wide ResNet-101-2 model.
    The model is the same as ResNet except for the bottleneck number of channels
    which is twice larger in every block. The number of channels in outer 1x1
    convolutions is the same.
    r•  rÊ   r²  Úwide_resnet101_2rt  ro  s      r,   rµ  rµ  �  s/   € ô œJ¨}ÈÔM€JÜÐ,¨jÑW¼DÀÑ<VÈvÑ<VÑWÐWr.   c           	      óT   — t        t        dd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z.Constructs a ResNet-50 model w/ GroupNorm
    r~  Ú	groupnorm)rÌ   rÍ   r9   Úresnet50_gnrt  ro  s      r,   r¸  r¸  ›  s.   € ô œJ¨|ÈÔT€JÜ˜-¨ÑR´t¸JÑ7QÈ&Ñ7QÑRÐRr.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z(Constructs a ResNeXt50-32x4d model.
    r~  rÇ   r|   ©rÌ   rÍ   r4   r5   Úresnext50_32x4drt  ro  s      r,   r»  r»  £  s2   € ô œJ¨|ÈÐXYÔZ€JÜÐ+¨ZÑV¼4À
Ñ;UÈfÑ;UÑVÐVr.   c           	      ó\   — t        t        dddddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )zVConstructs a ResNeXt50d-32x4d model. ResNext50 w/ deep stem & avg pool downsample
    r~  rÇ   r|   rÚ   T)rÌ   rÍ   r4   r5   rÑ   rÒ   r¢   Úresnext50d_32x4drt  ro  s      r,   r½  r½  «  s=   € ô Ü ¸BÈ1Ø °$ô8€Jô Ð,¨jÑW¼DÀÑ<VÈvÑ<VÑWÐWr.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z*Constructs a ResNeXt-101 32x4d model.
    r•  rÇ   r|   rº  Úresnext101_32x4drt  ro  s      r,   r¿  r¿  µ  ó2   € ô œJ¨}È"ÐYZÔ[€JÜÐ,¨jÑW¼DÀÑ<VÈvÑ<VÑWÐWr.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z*Constructs a ResNeXt-101 32x8d model.
    r•  rÇ   rØ   rº  Úresnext101_32x8drt  ro  s      r,   rÂ  rÂ  ½  rÀ  r.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z*Constructs a ResNeXt-101 32x16d model
    r•  rÇ   rÙ   rº  Úresnext101_32x16drt  ro  s      r,   rÄ  rÄ  Å  ó3   € ô œJ¨}È"ÐY[Ô\€JÜÐ-¨zÑX¼TÀ*Ñ=WÐPVÑ=WÑXÐXr.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z*Constructs a ResNeXt-101 32x32d model
    r•  rÇ   rº  Úresnext101_32x32drt  ro  s      r,   rÇ  rÇ  Í  rÅ  r.   c           	      óV   — t        t        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z)Constructs a ResNeXt101-64x4d model.
    r•  r0   r|   rº  Úresnext101_64x4drt  ro  s      r,   rÉ  rÉ  Õ  rÀ  r.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	z»Constructs an ECA-ResNeXt-26-T model.
    This is technically a 28 layer ResNet, like a 'D' bag-of-tricks model but with tiered 24, 32, 64 channels
    in the deep stem and ECA attn.
    rv  rÇ   rk  TÚeca©r:   ©rÌ   rÍ   rÑ   rÒ   r¢   rÖ   Úecaresnet26trt  ro  s      r,   rÎ  rÎ  Ý  s@   € ô Ü ¸"Ø¨$¼4È5Ô;QôS€Jô ˜.¨*ÑS¼¸ZÑ8RÈ6Ñ8RÑSÐSr.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	z-Constructs a ResNet-50-D model with eca.
    r~  rÇ   rÚ   TrË  rÌ  rÍ  Úecaresnet50drt  ro  s      r,   rÐ  rÐ  é  s@   € ô Ü ¸"ÈÐY]Ü 5Ô)ô+€Jô ˜.¨*ÑS¼¸ZÑ8RÈ6Ñ8RÑSÐSr.   c           
      ór   — t        t        ddddt        d¬«      ¬«      }t        d| fd	dit        |fi |¤Ž¤ŽS )
z…Constructs a ResNet-50-D model pruned with eca.
        The pruning has been obtained using https://arxiv.org/pdf/2002.08258.pdf
    r~  rÇ   rÚ   TrË  rÌ  rÍ  Úecaresnet50d_prunedÚprunedrt  ro  s      r,   rÒ  rÒ  ó  sH   € ô
 Ü ¸"ÈÐY]Ü 5Ô)ô+€Jô Ð/°ÑgÀDÐgÌDÐQ[ÑLfÐ_eÑLfÑgÐgr.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	z�Constructs an ECA-ResNet-50-T model.
    Like a 'D' bag-of-tricks model but with tiered 24, 32, 64 channels in the deep stem and ECA attn.
    r~  rÇ   rk  TrË  rÌ  rÍ  Úecaresnet50trt  ro  s      r,   rÕ  rÕ  þ  s@   € ô
 Ü ¸"Ø¨$¼4È5Ô;QôS€Jô ˜.¨*ÑS¼¸ZÑ8RÈ6Ñ8RÑSÐSr.   c           	      ól   — t        t        dddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z3Constructs a ResNet-50-D light model with eca.
    )r   r   é   rC   rÇ   TrË  rÌ  )rÌ   rÍ   rÑ   r¢   rÖ   Úecaresnetlightrt  ro  s      r,   rØ  rØ  	  s>   € ô Ü ¸2ÈÜ 5Ô)ô+€Jô Ð*¨JÑU¼$¸zÑ:TÈVÑ:TÑUÐUr.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	z.Constructs a ResNet-101-D model with eca.
    r•  rÇ   rÚ   TrË  rÌ  rÍ  Úecaresnet101drt  ro  s      r,   rÚ  rÚ    s@   € ô Ü ¸2ÈÐZ^Ü 5Ô)ô+€Jô ˜/¨:ÑT¼¸jÑ9SÈFÑ9SÑTÐTr.   c           
      ór   — t        t        ddddt        d¬«      ¬«      }t        d| fd	dit        |fi |¤Ž¤ŽS )
z…Constructs a ResNet-101-D model pruned with eca.
       The pruning has been obtained using https://arxiv.org/pdf/2002.08258.pdf
    r•  rÇ   rÚ   TrË  rÌ  rÍ  Úecaresnet101d_prunedrÓ  rt  ro  s      r,   rÜ  rÜ    sH   € ô
 Ü ¸2ÈÐZ^Ü 5Ô)ô+€Jô Ð0°*ÑhÀTÐhÌTÐR\ÑMgÐ`fÑMgÑhÐhr.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	z.Constructs a ResNet-200-D model with ECA.
    r¬  rÇ   rÚ   TrË  rÌ  rÍ  Úecaresnet200drt  ro  s      r,   rÞ  rÞ  (  ó@   € ô Ü ¸BÈ&Ð[_Ü 5Ô)ô+€Jô ˜/¨:ÑT¼¸jÑ9SÈFÑ9SÑTÐTr.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	z.Constructs a ResNet-269-D model with ECA.
    ©rC   é   é0   rØ   rÇ   rÚ   TrË  rÌ  rÍ  Úecaresnet269drt  ro  s      r,   rä  rä  2  rß  r.   c                 ór   — t        t        ddddddt        d¬«      ¬«      }t        d	| fi t        |fi |¤Ž¤ŽS )
zàConstructs an ECA-ResNeXt-26-T model.
    This is technically a 28 layer ResNet, like a 'D' bag-of-tricks model but with tiered 24, 32, 64 channels
    in the deep stem. This model replaces SE module with the ECA module
    rv  rÇ   r|   rk  TrË  rÌ  ©rÌ   rÍ   r4   r5   rÑ   rÒ   r¢   rÖ   Úecaresnext26t_32x4drt  ro  s      r,   rç  rç  <  óG   € ô Ü ¸2È!ÐXZØ¨$¼4È5Ô;QôS€Jô Ð/°ÑZ¼tÀJÑ?YÐRXÑ?YÑZÐZr.   c                 ór   — t        t        ddddddt        d¬«      ¬«      }t        d	| fi t        |fi |¤Ž¤ŽS )
zàConstructs an ECA-ResNeXt-50-T model.
    This is technically a 28 layer ResNet, like a 'D' bag-of-tricks model but with tiered 24, 32, 64 channels
    in the deep stem. This model replaces SE module with the ECA module
    rv  rÇ   r|   rk  TrË  rÌ  ræ  Úecaresnext50t_32x4drt  ro  s      r,   rê  rê  H  rè  r.   c           	      óh   — t        t        dt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nrv  rV   rÌ  ©rÌ   rÍ   rÖ   Ú
seresnet18rn  ro  s      r,   rí  rí  T  ó2   € äœJ¨|ÌÐX\ÔH]Ô^€JÜ˜,¨
ÑQ´d¸:Ñ6PÈÑ6PÑQÐQr.   c           	      óh   — t        t        dt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nr~  rV   rÌ  rì  Ú
seresnet34rn  ro  s      r,   rð  rð  Z  rî  r.   c           	      óh   — t        t        dt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nr~  rV   rÌ  rì  Ú
seresnet50rt  ro  s      r,   rò  rò  `  rî  r.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	Nr~  rÇ   rk  TrV   rÌ  rÍ  Úseresnet50trt  ro  s      r,   rô  rô  f  s=   € äÜ ¸2ÈØ¤$°$Ô"7ô9€Jô ˜-¨ÑR´t¸JÑ7QÈ&Ñ7QÑRÐRr.   c           	      óh   — t        t        dt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nr•  rV   rÌ  rì  Úseresnet101rt  ro  s      r,   rö  rö  n  ó2   € äœJ¨}ÌÐY]ÔI^Ô_€JÜ˜-¨ÑR´t¸JÑ7QÈ&Ñ7QÑRÐRr.   c           	      óh   — t        t        dt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nr¢  rV   rÌ  rì  Úseresnet152rt  ro  s      r,   rù  rù  t  r÷  r.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	Nr¢  rÇ   rÚ   TrV   rÌ  rÍ  Úseresnet152drt  ro  s      r,   rû  rû  z  s=   € äÜ ¸2ÈØ¤$°$Ô"7ô9€Jô ˜.¨*ÑS¼¸ZÑ8RÈ6Ñ8RÑSÐSr.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	z2Constructs a ResNet-200-D model with SE attn.
    r¬  rÇ   rÚ   TrV   rÌ  rÍ  Úseresnet200drt  ro  s      r,   rý  rý  ‚  ó?   € ô Ü ¸BÈ&Ø¤$°$Ô"7ô9€Jô ˜.¨*ÑS¼¸ZÑ8RÈ6Ñ8RÑSÐSr.   c           
      ón   — t        t        ddddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )	z2Constructs a ResNet-269-D model with SE attn.
    rá  rÇ   rÚ   TrV   rÌ  rÍ  Úseresnet269drt  ro  s      r,   r   r   Œ  rþ  r.   c                 ór   — t        t        ddddddt        d¬«      ¬«      }t        d	| fi t        |fi |¤Ž¤ŽS )
zÃConstructs a SE-ResNeXt-26-D model.`
    This is technically a 28 layer ResNet, using the 'D' modifier from Gluon / bag-of-tricks for
    combination of deep stem and avg_pool in downsample.
    rv  rÇ   r|   rÚ   TrV   rÌ  ræ  Úseresnext26d_32x4drt  ro  s      r,   r  r  –  sG   € ô Ü ¸2È!ÐXZØ 4´DÀDÔ4IôK€Jô Ð.°
ÑY¼dÀ:Ñ>XÐQWÑ>XÑYÐYr.   c                 ór   — t        t        ddddddt        d¬«      ¬«      }t        d	| fi t        |fi |¤Ž¤ŽS )
z«Constructs a SE-ResNet-26-T model.
    This is technically a 28 layer ResNet, like a 'D' bag-of-tricks model but with tiered 24, 32, 64 channels
    in the deep stem.
    rv  rÇ   r|   rk  TrV   rÌ  ræ  Úseresnext26t_32x4drt  ro  s      r,   r  r  ¢  sG   € ô Ü ¸2È!ÐXZØ¨$¼4È4Ô;PôR€Jô Ð.°
ÑY¼dÀ:Ñ>XÐQWÑ>XÑYÐYr.   c           	      ól   — t        t        dddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nr~  rÇ   r|   rV   rÌ  ©rÌ   rÍ   r4   r5   rÖ   Úseresnext50_32x4drt  ro  s      r,   r  r  ®  s=   € äÜ ¸2È!Ü 4Ô(ô*€Jô Ð-¨zÑX¼TÀ*Ñ=WÐPVÑ=WÑXÐXr.   c           	      ól   — t        t        dddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nr•  rÇ   r|   rV   rÌ  r  Úseresnext101_32x4drt  ro  s      r,   r	  r	  ¶  ó=   € äÜ ¸BÈ1Ü 4Ô(ô*€Jô Ð.°
ÑY¼dÀ:Ñ>XÐQWÑ>XÑYÐYr.   c           	      ól   — t        t        dddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nr•  rÇ   rØ   rV   rÌ  r  Úseresnext101_32x8drt  ro  s      r,   r  r  ¾  r
  r.   c                 ór   — t        t        ddddddt        d¬«      ¬«      }t        d	| fi t        |fi |¤Ž¤ŽS )
Nr•  rÇ   rØ   rÚ   TrV   rÌ  ræ  Úseresnext101d_32x8drt  ro  s      r,   r  r  Æ  sD   € äÜ ¸BÈ1Ø °$Ü 4Ô(ô*€Jô Ð/°ÑZ¼tÀJÑ?YÐRXÑ?YÑZÐZr.   c           	      ól   — t        t        dddt        d¬«      ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )Nr•  r0   r|   rV   rÌ  r  Úseresnext101_64x4drt  ro  s      r,   r  r  Ï  r
  r.   c                 ór   — t        t        ddddddt        d¬«      ¬	«      }t        d
| fi t        |fi |¤Ž¤ŽS )Nr¢  r0   r|   rÚ   rC   r)   rV   rÌ  )rÌ   rÍ   r4   r5   rÒ   r¡   rÔ   rÖ   Úsenet154rt  ro  s      r,   r  r  ×  sC   € äÜ ¸BÈ1ÐX^Ø¨q¼TÈTÔ=RôT€Jô ˜* jÑO´D¸Ñ4NÀvÑ4NÑOÐOr.   c           	      ó\   — t        t        dt        ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z9Constructs a ResNet-18 model with blur anti-aliasing
    rv  ©rÌ   rÍ   r;   Úresnetblur18)r¯   r"   r   r4  ro  s      r,   r  r  ß  ó.   € ô œJ¨|ÄjÔQ€JÜ˜.¨*ÑS¼¸ZÑ8RÈ6Ñ8RÑSÐSr.   c           	      ó\   — t        t        dt        ¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z9Constructs a ResNet-50 model with blur anti-aliasing
    r~  r  Úresnetblur50©r¯   r#   r   r4  ro  s      r,   r  r  ç  r  r.   c           	      ób   — t        t        dt        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z;Constructs a ResNet-50-D model with blur anti-aliasing
    r~  rÇ   rÚ   T©rÌ   rÍ   r;   rÑ   rÒ   r¢   Úresnetblur50dr  ro  s      r,   r  r  ï  s:   € ô Ü ¼
Ø °$ô8€Jô ˜/¨:ÑT¼¸jÑ9SÈFÑ9SÑTÐTr.   c           	      ób   — t        t        dt        ddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z<Constructs a ResNet-101-D model with blur anti-aliasing
    r•  rÇ   rÚ   Tr  Úresnetblur101dr  ro  s      r,   r  r  ù  s;   € ô Ü ¼Ø °$ô8€Jô Ð*¨JÑU¼$¸zÑ:TÈVÑ:TÑUÐUr.   c           	      óv   — t        t        dt        j                  ddd¬«      }t	        d| fi t        |fi |¤Ž¤ŽS )z<Constructs a ResNet-34-D model w/ avgpool anti-aliasing
    r~  rÇ   rÚ   Tr  Úresnetaa34d)r¯   r"   rM   r‘   r4  ro  s      r,   r   r     s?   € ô Ü ¼¿¹ÐRTÐ`fÐquôw€Jä˜-¨ÑR´t¸JÑ7QÈ&Ñ7QÑRÐRr.   c           	      óp   — t        t        dt        j                  ¬«      }t	        d| fi t        |fi |¤Ž¤ŽS )z<Constructs a ResNet-50 model with avgpool anti-aliasing
    r~  r  Ú
resnetaa50©r¯   r#   rM   r‘   r4  ro  s      r,   r"  r"    s2   € ô œJ¨|ÄbÇlÁlÔS€JÜ˜,¨
ÑQ´d¸:Ñ6PÈÑ6PÑQÐQr.   c           	      óv   — t        t        dt        j                  ddd¬«      }t	        d| fi t        |fi |¤Ž¤ŽS )z>Constructs a ResNet-50-D model with avgpool anti-aliasing
    r~  rÇ   rÚ   Tr  Úresnetaa50dr#  ro  s      r,   r%  r%    s>   € ô Ü ¼¿¹Ø °$ô8€Jô ˜-¨ÑR´t¸JÑ7QÈ&Ñ7QÑRÐRr.   c           	      óv   — t        t        dt        j                  ddd¬«      }t	        d| fi t        |fi |¤Ž¤ŽS )z?Constructs a ResNet-101-D model with avgpool anti-aliasing
    r•  rÇ   rÚ   Tr  Úresnetaa101dr#  ro  s      r,   r'  r'    s>   € ô Ü ¼¿¹Ø °$ô8€Jô ˜.¨*ÑS¼¸ZÑ8RÈ6Ñ8RÑSÐSr.   c                 óŒ   — t        t        dt        j                  dddt        d¬«      ¬«      }t	        d| fi t        |fi |¤Ž¤ŽS )	zAConstructs a SE=ResNet-50-D model with avgpool anti-aliasing
    r~  rÇ   rÚ   TrV   rÌ  )rÌ   rÍ   r;   rÑ   rÒ   r¢   rÖ   Úseresnetaa50dr#  ro  s      r,   r)  r)  (  sG   € ô Ü ¼¿¹Ø °$Ä4ÐSWÔCXôZ€Jô ˜/¨:ÑT¼¸jÑ9SÈFÑ9SÑTÐTr.   c                 ó�   — t        t        ddddddt        j                  t        d¬«      ¬«	      }t	        d	| fi t        |fi |¤Ž¤ŽS )
úIConstructs a SE=ResNeXt-101-D 32x8d model with avgpool anti-aliasing
    r•  rÇ   rØ   rÚ   TrV   rÌ  ©	rÌ   rÍ   r4   r5   rÑ   rÒ   r¢   r;   rÖ   Úseresnextaa101d_32x8dr#  ro  s      r,   r-  r-  2  sL   € ô Ü ¸BÈ1Ø °$ÄÇÁÜ 4Ô(ô*€Jô Ð1°:Ñ\ÄÀjÑA[ÐTZÑA[Ñ\Ð\r.   c                 ó�   — t        t        ddddddt        j                  t        d¬«      ¬	«	      }t	        d
| fi t        |fi |¤Ž¤ŽS )r+  )rC   r­  r£  r|   rÇ   rØ   r0   rÚ   TrV   rÌ  r,  Úseresnextaa201d_32x8dr#  ro  s      r,   r/  r/  =  sL   € ô Ü ¸RÈAØ °$ÄÇÁÜ 4Ô(ô*€Jô Ð1°:Ñ\ÄÀjÑA[ÐTZÑA[Ñ\Ð\r.   c                 óœ   — t        t        d«      d¬«      }t        t        dddddt        |¬«      ¬	«      }t	        d
| fi t        |fi |¤Ž¤ŽS )zÓConstructs a ResNet-RS-50 model.
    Paper: Revisiting ResNets - https://arxiv.org/abs/2103.07579
    Pretrained weights from https://github.com/tensorflow/tpu/tree/bee9c4f6/models/official/resnet/resnet_rs
    rV   r™   ©Úrd_ratior~  rÇ   rÚ   TrÌ  ©rÌ   rÍ   rÑ   rÒ   rÓ   r¢   rÖ   Ú
resnetrs50©r   r   r¯   r#   r4  ©r2  r³   r:   rp  s       r,   r4  r4  H  sS   € ô œ $›°$Ô7€JÜÜ ¸"ÈÐbfØ¤4°:Ô#>ô@€Jô ˜,¨
ÑQ´d¸:Ñ6PÈÑ6PÑQÐQr.   c                 óœ   — t        t        d«      d¬«      }t        t        dddddt        |¬«      ¬	«      }t	        d
| fi t        |fi |¤Ž¤ŽS )zÔConstructs a ResNet-RS-101 model.
    Paper: Revisiting ResNets - https://arxiv.org/abs/2103.07579
    Pretrained weights from https://github.com/tensorflow/tpu/tree/bee9c4f6/models/official/resnet/resnet_rs
    rV   r™   r1  r•  rÇ   rÚ   TrÌ  r3  Úresnetrs101r5  r6  s       r,   r8  r8  U  óS   € ô œ $›°$Ô7€JÜÜ ¸2ÈÐcgØ¤4°:Ô#>ô@€Jô ˜-¨ÑR´t¸JÑ7QÈ&Ñ7QÑRÐRr.   c                 óœ   — t        t        d«      d¬«      }t        t        dddddt        |¬«      ¬	«      }t	        d
| fi t        |fi |¤Ž¤ŽS )zÔConstructs a ResNet-RS-152 model.
    Paper: Revisiting ResNets - https://arxiv.org/abs/2103.07579
    Pretrained weights from https://github.com/tensorflow/tpu/tree/bee9c4f6/models/official/resnet/resnet_rs
    rV   r™   r1  r¢  rÇ   rÚ   TrÌ  r3  Úresnetrs152r5  r6  s       r,   r;  r;  b  r9  r.   c                 óœ   — t        t        d«      d¬«      }t        t        dddddt        |¬«      ¬	«      }t	        d
| fi t        |fi |¤Ž¤ŽS )zÔConstructs a ResNet-RS-200 model.
    Paper: Revisiting ResNets - https://arxiv.org/abs/2103.07579
    Pretrained weights from https://github.com/tensorflow/tpu/tree/bee9c4f6/models/official/resnet/resnet_rs
    rV   r™   r1  r¬  rÇ   rÚ   TrÌ  r3  Úresnetrs200r5  r6  s       r,   r=  r=  o  óS   € ô œ $›°$Ô7€JÜÜ ¸BÈ&ÐdhØ¤4°:Ô#>ô@€Jô ˜-¨ÑR´t¸JÑ7QÈ&Ñ7QÑRÐRr.   c                 óœ   — t        t        d«      d¬«      }t        t        dddddt        |¬«      ¬	«      }t	        d
| fi t        |fi |¤Ž¤ŽS )zÔConstructs a ResNet-RS-270 model.
    Paper: Revisiting ResNets - https://arxiv.org/abs/2103.07579
    Pretrained weights from https://github.com/tensorflow/tpu/tree/bee9c4f6/models/official/resnet/resnet_rs
    rV   r™   r1  )r|   é   é5   r|   rÇ   rÚ   TrÌ  r3  Úresnetrs270r5  r6  s       r,   rB  rB  |  r>  r.   c                 óœ   — t        t        d«      d¬«      }t        t        dddddt        |¬«      ¬	«      }t	        d
| fi t        |fi |¤Ž¤ŽS )zÔConstructs a ResNet-RS-350 model.
    Paper: Revisiting ResNets - https://arxiv.org/abs/2103.07579
    Pretrained weights from https://github.com/tensorflow/tpu/tree/bee9c4f6/models/official/resnet/resnet_rs
    rV   r™   r1  )r|   r£  éH   r|   rÇ   rÚ   TrÌ  r3  Úresnetrs350r5  r6  s       r,   rE  rE  Š  r>  r.   c                 óœ   — t        t        d«      d¬«      }t        t        dddddt        |¬«      ¬	«      }t	        d
| fi t        |fi |¤Ž¤ŽS )zÓConstructs a ResNet-RS-420 model
    Paper: Revisiting ResNets - https://arxiv.org/abs/2103.07579
    Pretrained weights from https://github.com/tensorflow/tpu/tree/bee9c4f6/models/official/resnet/resnet_rs
    rV   r™   r1  )r|   é,   éW   r|   rÇ   rÚ   TrÌ  r3  Úresnetrs420r5  r6  s       r,   rI  rI  —  r>  r.   c           	      óz   — t        t        t        t        t        gddddd¬«      }t        d| fi t        |fi |¤Ž¤ŽS )z)Constructs a tiny ResNet test model.
    rj  rÙ   rÚ   T)rÇ   rã  rã  é`   )rÌ   rÍ   rÑ   rÒ   r¢   rH   Útest_resnet)r¯   r"   r#   r4  ro  s      r,   rL  rL  ¤  sE   € ô Üœ:¤z´:Ð>À|Ø °$ÐAQôS€Jô ˜-¨ÑR´t¸JÑ7QÈ&Ñ7QÑRÐRr.   Útv_resnet34Útv_resnet50Útv_resnet101Útv_resnet152Útv_resnext50_32x4dÚig_resnext101_32x8dÚig_resnext101_32x16dÚig_resnext101_32x32dÚig_resnext101_32x48dÚssl_resnet18Ússl_resnet50Ússl_resnext50_32x4dÚssl_resnext101_32x4dÚssl_resnext101_32x8dÚssl_resnext101_32x16dÚswsl_resnet18Úswsl_resnet50Úswsl_resnext50_32x4dÚswsl_resnext101_32x4dÚswsl_resnext101_32x8dÚswsl_resnext101_32x16dÚgluon_resnet18_v1bÚgluon_resnet34_v1bÚgluon_resnet50_v1bÚgluon_resnet101_v1bÚgluon_resnet152_v1bÚgluon_resnet50_v1cÚgluon_resnet101_v1cÚgluon_resnet152_v1cÚgluon_resnet50_v1dÚgluon_resnet101_v1dÚgluon_resnet152_v1dÚgluon_resnet50_v1sÚgluon_resnet101_v1sr  )	Úgluon_resnet152_v1sÚgluon_resnext50_32x4dÚgluon_resnext101_32x4dÚgluon_resnext101_64x4dÚgluon_seresnext50_32x4dÚgluon_seresnext101_32x4dÚgluon_seresnext101_64x4dÚgluon_senet154Úseresnext26tn_32x4d)r   )r   r   NNNN)r§   )r   rÇ   r   Fr§   r§   NNr-  )rÉ   )”rr   r   Ú	functoolsr   Útypingr   r   r   r   r   r	   r
   rw   Útorch.nnrM   Útorch.nn.functionalÚ
functionalr(  Ú	timm.datar   r   Útimm.layersr   r   r   r   r   r   r   r   r   r   r   r   Ú_builderr   Ú	_featuresr   Ú_manipulater   Ú	_registryr   r   r    Ú__all__ru   r-   rv   r"   r#   rŒ   r•   r*  r�   r/  r.  rÄ   r!   r4  rD  rH  rQ  rU  rZ  r\  Údefault_cfgsrm  rs  rx  r{  r  r�  rƒ  r…  r‡  r‰  rŒ  r�  r‘  r“  r—  rš  r�  r   r¤  r¦  r¨  rª  r®  r°  r³  rµ  r¸  r»  r½  r¿  rÂ  rÄ  rÇ  rÉ  rÎ  rÐ  rÒ  rÕ  rØ  rÚ  rÜ  rÞ  rä  rç  rê  rí  rð  rò  rô  rö  rù  rû  rý  r   r  r  r  r	  r  r  r  r  r  r  r  r  r   r"  r%  r'  r)  r-  r/  r4  r8  r;  r=  rB  rE  rI  rL  ro   r*   r.   r,   ú<module>r…     s$&  ðñó Ý ß @× @Ñ @ã Ý ß Ð ç A÷U÷ U÷ Uó Uå *Ý +Ý 'ß YÑ Yâ
0€ò˜Sð ¨#ð ¸ð ÀSó ô
u�—‘ô uôpu�—‘ô uðx ØØ(,Ø04ØØòØðàðð ðð ð	ð
 ðð ! ™ðð ˜T "§)¡)™_Ñ-ðð ‡Y�YóðH ØØ(,Ø04ØØòØðàðð ðð ð	ð
 ðð ! ™ðð ˜T "§)¡)™_Ñ-ðð ‡Y�Yóò6j˜5ð j¨$¨x¸Ñ/@Ñ*Aó jð( ØØ !ØØ!#Ø "ØØòV Ø˜˜t JÑ/°°jÑ1AÐAÑBÀCÐGÑHðV à˜˜S˜‘/ðV ð ˜S #˜X‘ðV ð ð	V ð
 ðV ð ðV ð ðV ð ðV ð ðV ð ðV ð ˆ4��c˜2Ÿ9™9�nÑ%Ñ&¨¨T°#°s°(©^Ñ(<Ð<Ñ=óV ôrUˆR�Y‰Yô Uòp
G˜Cð G¨Tð GÈó Gò
ˆcð 
 T¨#¨s¨(¡^ó 
òIˆsð I d¨3°¨8¡nó Iò
�ð  t¨C°¨H¡~ó òˆsð  d¨3°¨8¡nó ò�ð  t¨C°¨H¡~ó òˆsð  d¨3°¨8¡nó ñ %ð s	&à™Øð AØ ¨FÀ$ÐXeØô	ðs	&ð ™Øð AØ ¨FÀ$ÐXeØô	ðs	&ð ™ØØôAðs	&ð ™ØØôAðs	&ð$ ™ØØôAð%s	&ð* ™&ØØyØôð+s	&ð2 $¡UØØ /¸CÈIô&Wð3s	&ð8 ™ØØôAð9s	&ð> ™ØØôAð?s	&ðD ™ØØØôðEs	&ðL ™ØØtôvðMs	&ðR #¡EØØ /¸Cô%AðSs	&ðX ™&ØØyØôðYs	&ð` ™ØØtôvðas	&ðf ™ØØuØôðgs	&ðn ™&Øð CØ¨À&Ø }ÀCô	Iðos	&ðx ™ØØ}ôñys	&ð~ ™Øð DØ ¨F¸SÐR_Ðorôtðs	&ðF ™ØØôAðGs	&ðL ™ØØôAðMs	&ðR ™ØØ~ô@ðSs	&ðX ™ØØ~ô@ðYs	&ð^ ™ØØ}ôð_s	&ðd ™ØØ}ôðes	&ðj ‘uØØ|ô~ðks	&ðp ™ØØzô|ðqs	&ðv ™ØØyô{ðws	&ð| ™ØØyô{ð}s	&ðB ™ØØyô{ðCs	&ðH ™&ØØyØôðIs	&ðP $¡UØØ /Ø }ÀCØô	&ðQs	&ðZ ™Øð AØôð[s	&ðb ™Øð AØôðcs	&ðj ™Øð AØôòks	&ðr ™6¨YÔ7ðss	&ðt ™%ØØ}ôðus	&ðz ™Øð AôBð{s	&ð@ ™Øð AôBðAs	&ðF ™Øð AôBðGs	&ðL ™6ØØzØ¨À&ÐSWØ¨=ô	:ðMs	&ðV ™%ØØ}ôðWs	&ð\ ™Øð AôBð]s	&ðb ™Øð AôBðcs	&ðh ™Øð AôBðis	&ðn ™6ØØzØ¨À&ÐSWØ¨=ô	:ðos	&ðx ™6›8ðys	&ðz ™6ØØzØ¨À&ÐSWØ¨=ô	:ð{s	&ðD  ¡ØØ~ô"@ðEs	&ðN ™ØØGØÐ+NôPðOs	&ðV ™ØØGØÐ+NôPðWs	&ð^ ™ØØGØÐ+NôPò_s	&ðf ™ØØGØ ¨FÀMÐafØÐ+Nô	Pðgs	&ðp ™ØØHØÐ+NôPðqs	&ðx ™$ØØHØ ¨FÀMÐafØÐ+Nô	Pðys	&ðB ™ØØHØÐ+NôPðCs	&ðJ ™$ØØHØ ¨FÀMÐafØÐ+Nô	PðKs	&ðT ™tØØNØÐ+Nô PðUs	&ð\ ¡ØØNØ ¨FÀMÐafØÐ+Nô	!Pð]s	&ðf ¡ØØOØÐ+Nô!Pðgs	&ðn  ¡ØØOØ ¨FÀMÐafØÐ+Nô	"Pðos	&ð| ™FØð AØôð}s	&ðH ¡Øð Dô!EðIs	&ðN ™uØð Gô HðOs	&ðT ™uØð Gô HðUs	&ðZ ™vØð Gô Hð[s	&ð` ™vØð Aô Bðas	&ðf ¡ØØ|Øô!ðgs	&ðn !¡&£(òos	&ðp ¡Øð Cô!Dðqs	&ðz ™tØØNØÐ+Nô Pð{s	&ðB ¡ØØOØÐ+Nô!PðCs	&ðJ ¡ØØOØÐ+Nô!PðKs	&ðR ¡ØØNØ ¨FÀMÐafØÐ+Nô	!PðSs	&ð\  ¡ØØOØ ¨FÀMÐafØÐ+Nô	"Pð]s	&ðn +©DØØQØÐ+[ô-]ðos	&ðv ,©TØØRØÐ+[ô.]ðws	&ð~ ,©TØØRØÐ+[ô.]ðs	&ðF	 ,©TØØRØÐ+[ô.]ðG	s	&ðT	 '©ØØrØÐ+rô*tðU	s	&ð\	 '©ØØrØÐ+rô*tð]	s	&ðd	 .©tØØxØÐ+rô0tðe	s	&ðl	 /±ØØyØÐ+rô1tðm	s	&ðt	 /±ØØyØÐ+rô1tðu	s	&ð|	 0±ØØzØÐ+rô2tð}	s	&ðJ
 $¡TØØyØÐ+rô&tòK
s	&ðR
 $¡TØØyØÐ+rô&tðS
s	&ðZ
 +©DØØØÐ+rô-tð[
s	&ðb
 ,©TØñ AØÐ+rô.tðc
s	&ñj
 ,©TØñ AØÐ+rô.tðk
s	&ñr
 -©dØñ BØÐ+rô/tðs
s	&ñ~
 ™VØÙ|Ø¨À&Ø¨Mõ	;ð
s	&ñH ¡ØÙzØ¨Mõ!;ðIs	&ñP ™eØÙxØ¨DÀ-õQðQs	&ñX $¡UØÙzØ¨DÀ-õ&QðYs	&ñ` ™UØÙ|Ø¨À&Ø¨Mõ	;ðas	&ñj ™EØñ DØôðks	&ñr ™EØñ DØôðss	&ñz ™FØñ DØôð{s	&ñB ™uØÙyØ¨DÀ-õ QðCs	&ñJ %¡eØÙ{Ø¨DÀ-õ'QðKs	&ñR ™vØ¨ÀÐQWõ YðSs	&ñV ™fØñ BØ¨Á(ÐUYØ©=ô	:òWs	&ñd $¡U°iÔ%@ðes	&ñf $¡U°iÔ%@ðgs	&ñl ™F›Hðms	&ñn ™F›Hðos	&ñp ™%Øñ BØôðqs	&ñx ™%Øñ BØôðys	&ñ@ ™&Øñ BØôðAs	&ñH ™6ØÙ}ôðIs	&ñN ™VØôðOs	&ñR ™V›XðSs	&ñT ™V›XðUs	&ñV ™VØÙ|Ø¨À&ÐSWØ¨=ô	ðWs	&ñb ™fØ¨À&õJðcs	&ñf ™fØ¨À&õJðgs	&ñn !¡&ØÙ~Øô#ðos	&ñv !¡&ØÙØô#ðws	&ñ~ "¡6Øñ Cô$Dòs	&ñD #¡F£HðEs	&ñF !¡%Øñ Fô#GðGs	&ñL "¡5Øñ GØô$ðMs	&ñX #¡FØØ t¸3õ%@ðYs	&ñ^ $¡VØØ t¸3õ&@ð_s	&ñd 1±&ØØ -¹6ÐS`ÐpsØõ3ðes	&ñl -©fØØ¨Cõ/1ðms	&ñr 1±$ØÙ¨IÁÑVcÐnqõ3sðss	&ñx %¡dØÙ©¸yØ -¹8ÑUbÐruõ'wðys	&ñB ™FØÙ i¸$ÈcõSðCs	&ñH ™6ØÙ i¸$ÈcõSðIs	&ñN ™VØÙ i¸$ÈcõSðOs	&ñT %¡fØÙ i¸$Ècõ'SðUs	&ñ\ ™f›hð]s	&ñ^ ™FØÙxôzð_s	&ñd ™v°Ô;ðes	&ñf ¡°)Ô <ògs	&ñh ™V¨yÔ9ðis	&ñj ™5ØÙ~ô@ðks	&ñr ™v°Ô;ðss	&ñt $¡UØñ IØô&ðus	&ñ@ ™$ØÙ}Ù ©F¹TÐS`Ù¨Iõ	7ðAs	&ñJ ™4Øñ DÙ ¨F¸TÐS`Ù¨Iõ	7ðKs	&ñT ™4Øñ DØ ¨F¸SÐR_Ù¨Iõ	7ðUs	&ñ^ ™4ØÙ}Ø ¨F¸SÐR_Ù¨Iõ	7ð_s	&ñh ™4ØÙ~Ø ¨F¸SÑR_Ù¨Iõ	7ðis	&ñr ™4Øñ DØ ©F¸SÑR_Ù¨Iõ	7ðss	&ñ| ™4ØÙ~Ø ©H¸sÑTaÙ¨Iõ	7ð}s	&ñJ ™5Øñ AôBðKs	&ñP ™5Øñ AôBðQs	&ñV ™5Øñ AôBðWs	&ñ\ ™EØñ BôCð]s	&ñb ™EØñ BôCðcs	&ñh ™EØñ AØôòis	&ñp ™UØñ BØôðqs	&ñx ™UØñ BØôðys	&ñ@ ™EØñ AØôðAs	&ñH ™UØñ BØôðIs	&ñP ™UØñ BØôðQs	&ñX ™EØñ AØôðYs	&ñ` ™UØñ BØôðas	&ñh ™UØñ BØôðis	&ñp !¡%Øñ Dô#Eðqs	&ñv "¡5Øñ Eô$Fðws	&ñ| "¡5Øñ Eô$Fð}s	&ñB #¡EØñ Fô%GðCs	&ñH $¡UØñ Gô&HðIs	&ñN $¡UØñ Gô&HðOs	&ñT ™5ØÙ|ØôðUs	&ñ^ ™TØØ /¸DÙ ©F¸yõJñ_s	&ó s	€ðl òQ˜$ð Q°Vó Qó ðQð òQ˜$ð Q°Vó Qó ðQð òP˜ð P°Fó Pó ðPð òQ˜$ð Q°Vó Qó ðQð òP˜ð P°Fó Pó ðPð òQ˜$ð Q°Vó Qó ðQð òP˜ð P°Fó Pó ðPð òQ˜$ð Q°Vó Qó ðQð òQ˜$ð Q°Vó Qó ðQð òP˜ð P°Fó Pó ðPð òQ˜$ð Q°Vó Qó ðQð òQ˜$ð Q°Vó Qó ðQð òQ˜$ð Q°Vó Qó ðQð òQ˜$ð Q°Vó Qó ðQð òQ˜$ð Q°Vó Qó ðQð òR˜4ð R°fó Ró ðRð òR˜4ð R°fó Ró ðRð òR˜4ð R°fó Ró ðRð òQ˜$ð Q°Vó Qó ðQð òR˜4ð R°fó Ró ðRð òR˜4ð R°fó Ró ðRð òR˜4ð R°fó Ró ðRð òQ˜$ð Q°Vó Qó ðQð òR˜4ð R°fó Ró ðRð òW ð W¸6ó Wó ðWð òX ð X¸Fó Xó ðXð òS˜Dð S°vó Só ðSð òW ð W¸6ó Wó ðWð òX ð X¸Fó Xó ðXð òX ð X¸Fó Xó ðXð òX ð X¸Fó Xó ðXð òY $ð Y¸Vó Yó ðYð òY $ð Y¸Vó Yó ðYð òX ð X¸Fó Xó ðXð òT˜Tð T¸ó Tó ðTð òT˜Tð T¸ó Tó ðTð òh Dð h¸vó hó ðhð òT˜Tð T¸ó Tó ðTð òV˜tð V¸&ó Vó ðVð òU˜dð U¸ó Uó ðUð òi Tð iÀó ió ðið òU˜dð U¸ó Uó ðUð òU˜dð U¸ó Uó ðUð ò[ Dð [¸vó [ó ð[ð ò[ Dð [¸vó [ó ð[ð òR˜4ð R°fó Ró ðRð
 òR˜4ð R°fó Ró ðRð
 òR˜4ð R°fó Ró ðRð
 òS˜Dð S°vó Só ðSð òS˜Dð S°vó Só ðSð
 òS˜Dð S°vó Só ðSð
 òT˜Tð T¸ó Tó ðTð òT˜Tð T¸ó Tó ðTð òT˜Tð T¸ó Tó ðTð òZ 4ð Z¸fó Zó ðZð òZ 4ð Z¸fó Zó ðZð òY $ð Y¸Vó Yó ðYð òZ 4ð Z¸fó Zó ðZð òZ 4ð Z¸fó Zó ðZð ò[ Dð [¸vó [ó ð[ð òZ 4ð Z¸fó Zó ðZð òP˜ð P°Fó Pó ðPð òT˜Tð T¸ó Tó ðTð òT˜Tð T¸ó Tó ðTð òU˜dð U¸ó Uó ðUð òV˜tð V¸&ó Vó ðVð òS˜Dð S°vó Só ðSð òR˜4ð R°fó Ró ðRð òS˜Dð S°vó Só ðSð òT˜Tð T¸ó Tó ðTð òU˜dð U¸ó Uó ðUð ò] dð ]Àó ]ó ð]ð ò] dó ]ó ð]ð ò	R˜4ð 	R°fó 	Ró ð	Rð ò	S˜Dð 	S°vó 	Só ð	Sð ò	S˜Dð 	S°vó 	Só ð	Sð ò	S˜Dð 	S°vó 	Só ð	Sð ò	S˜Dð 	S°vó 	Só ð	Sð ò	S˜Dð 	S°vó 	Só ð	Sð ò	S˜Dð 	S°vó 	Só ð	Sð òS˜Dð S°vó Só ðSñ ˜Hð ,'ÙÐ%ð,'áÐ%ð,'ñ Ð'ð,'ñ Ð'ð	,'ñ
 Ð4ð,'ñ ÐAð,'ñ ÐBð,'ñ ÐBð,'ñ ÐBð,'ñ Ð6ð,'ñ Ð6ð,'ñ ÐDð,'ñ ÐFð,'ñ ÐFð,'ñ ÐHð,'ñ  Ð4ð!,'ñ" Ð4ñ#,'ñ$ ÐBð%,'ñ& ÐDð','ñ( ÑDð),'ñ* ÑFð+,'ñ, Ñ/ð-,'ñ. Ñ/ð/,'ñ0 Ñ/ð1,'ñ2 Ñ1ð3,'ñ4 Ñ1ð5,'ñ6 Ñ0ð7,'ñ8 Ñ2ð9,'ñ: Ñ2ð;,'ñ< Ñ0ð=,'ñ> Ñ2ð?,'ñ@ Ñ2ðA,'ñB Ñ0ðC,'ñD Ñ2ñE,'ñF 3Ù9Ù;Ù;Ù=Ù ?Ù ?Ù+Ù/óW,'õ ,r.   