§
    ·¯¸jçq  ã                   ó.  — 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
 ddlZddlmZ ddlmc mZ ddlmZmZ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"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. ddl/m0Z0m1Z1 ddl2m3Z3m4Z4 dgZ5 ej6        e7¦  «        Z8de9de9dej:        dej:        fd„Z; e.e;¦  «         dej:        dej:        dej:        dee9e9f         dee9e9f         dej:        fd„Z< G d„ dej=        ¦  «        Z> G d„ dej=        ¦  «        Z?dej:        de9deej:        ee9e9f         f         fd„Z@	 d:dej:        de9d ee9e9f         d!eee9e9f                  dej:        f
d"„ZA G d#„ dej=        ¦  «        ZBd$„ ZCd;d&„ZD e3 eDd'd(d)eedd*d+¬,¦  «         eDd-d(d)eedd*d+¬,¦  «         eDd.d(d)eedd*d+¬,¦  «         eDeed/d0d1¬2¦  «        d3œ¦  «        ZEd<d5„ZFe4d<deBfd6„¦   «         ZGe4d<deBfd7„¦   «         ZHe4d<deBfd8„¦   «         ZIe4d<deBfd9„¦   «         ZJdS )=a+   Vision Transformer (ViT) in PyTorch

A PyTorch implement of Vision Transformers as described in:

'Exploring Plain Vision Transformer Backbones for Object Detection'
    - https://arxiv.org/abs/2203.16527

'Segment Anything Model (SAM)'
    - https://github.com/facebookresearch/segment-anything/

é    N)Úpartial)ÚCallableÚListÚOptionalÚTupleÚTypeÚUnion)ÚIMAGENET_DEFAULT_MEANÚIMAGENET_DEFAULT_STDÚIMAGENET_INCEPTION_MEANÚIMAGENET_INCEPTION_STD)Úget_device_dtypeÚ
PatchEmbedÚMlpÚDropPathÚcalculate_drop_path_ratesÚPatchDropoutÚLayerNorm2dÚ
LayerScaleÚClassifierHeadÚNormMlpClassifierHeadÚFormatÚresample_abs_pos_embed_nhwcÚRotaryEmbeddingCatÚapply_rot_embed_catÚ	to_2tupleÚuse_fused_attn)ÚFinalé   )Úbuild_model_with_cfg)Úfeature_take_indices)Úregister_notrace_function)Ú
checkpointÚcheckpoint_seq)Úgenerate_default_cfgsÚregister_modelÚVisionTransformerSAMÚq_sizeÚk_sizeÚrel_posÚreturnc                 óÆ  — t          dt          | |¦  «        z  dz
  ¦  «        }|j        d         |k    rwt          j        |                     d|j        d         d¦  «                             ddd¦  «        |d¬¦  «        }|                     d|¦  «                             dd¦  «        }n|}t          j        | t          j	        ¬¦  «        dd…df         t          || z  d	¦  «        z  }t          j        |t          j	        ¬¦  «        ddd…f         t          | |z  d	¦  «        z  }||z
  |dz
  t          | |z  d	¦  «        z  z   }|| 
                    ¦   «                  S )
a\  
    Get relative positional embeddings according to the relative positions of
        query and key sizes.
    Args:
        q_size (int): size of query q.
        k_size (int): size of key k.
        rel_pos (Tensor): relative position embeddings (L, C).

    Returns:
        Extracted positional embeddings according to relative positions.
    é   r   r   éÿÿÿÿÚlinear)ÚsizeÚmode)ÚdtypeNç      ð?)ÚintÚmaxÚshapeÚFÚinterpolateÚreshapeÚpermuteÚtorchÚarangeÚfloat32Úlong)r(   r)   r*   Úmax_rel_distÚrel_pos_resizedÚq_coordsÚk_coordsÚrelative_coordss           úr/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/timm/models/vision_transformer_sam.pyÚget_rel_posrE   5   s]  € õ �q�3˜v vÑ.Ô.Ñ.°Ñ2Ñ3Ô3€Là„}�QÔ˜<Ò'Ð'åœ-Ø�OŠO˜A˜wœ}¨QÔ/°Ñ4Ô4×<Ò<¸QÀÀ1ÑEÔEØØð
ñ 
ô 
ˆð
 *×1Ò1°"°lÑCÔC×KÒKÈAÈqÑQÔQˆˆà!ˆõ Œ|˜F­%¬-Ð8Ñ8Ô8¸¸¸¸D¸ÔAÅCÈÐQWÉÐY\ÑD]ÔD]Ñ]€HÝŒ|˜F­%¬-Ð8Ñ8Ô8¸¸q¸q¸q¸ÔAÅCÈÐQWÉÐY\ÑD]ÔD]Ñ]€HØ (Ñ*¨v¸©z½SÀÈ&ÁÐRUÑ=VÔ=VÑ.VÑV€Oà˜?×/Ò/Ñ1Ô1Ô2Ð2ó    ÚqÚ	rel_pos_hÚ	rel_pos_wc                 óˆ  — |\  }}|\  }}t          |||¦  «        }	t          |||¦  «        }
| j        \  }}}|                      ||||¦  «        }t          j        d||	¦  «        }t          j        d||
¦  «        }|dd…dd…dd…dd…df         |dd…dd…dd…ddd…f         z   }|                     d||z  ||z  ¦  «        S )a£  
    Calculate decomposed Relative Positional Embeddings from :paper:`mvitv2`.
    https://github.com/facebookresearch/mvit/blob/19786631e330df9f3622e5402b4a419a263a2c80/mvit/models/attention.py
    Args:
        q (Tensor): query q in the attention layer with shape (B, q_h * q_w, C).
        rel_pos_h (Tensor): relative position embeddings (Lh, C) for height axis.
        rel_pos_w (Tensor): relative position embeddings (Lw, C) for width axis.
        q_size (Tuple): spatial sequence size of query q with (q_h, q_w).
        k_size (Tuple): spatial sequence size of key k with (k_h, k_w).

    Returns:
        bias (Tensor): attention bias to add to attention map
    zbhwc,hkc->bhwkzbhwc,wkc->bhwkNr.   )rE   r6   r9   r;   Úeinsum)rG   rH   rI   r(   r)   Úq_hÚq_wÚk_hÚk_wÚRhÚRwÚBÚ_ÚdimÚr_qÚrel_hÚrel_wÚ	attn_biass                     rD   Úget_decomposed_rel_pos_biasrY   X   sî   € ð( �H€CˆØ�H€CˆÝ	�S˜#˜yÑ	)Ô	)€BÝ	�S˜#˜yÑ	)Ô	)€Bà”�I€A€qˆ#Ø
�)Š)�A�s˜C Ñ
%Ô
%€CÝŒLÐ)¨3°Ñ3Ô3€EÝŒLÐ)¨3°Ñ3Ô3€Eà�a�a�a˜˜˜˜A˜A˜A˜q˜q˜q $Ð&Ô'¨%°°°°1°1°1°a°a°a¸¸q¸q¸qÐ0@Ô*AÑA€IØ×Ò˜R  s¡¨C°#©IÑ6Ô6Ð6rF   c                   óÐ   ‡ — e Zd ZU ee         ed<   dddddej        dddddfdeded	ed
ede	de	de
ej                 dedeeeef                  deej                 fˆ fd„Zd„ Zˆ xZS )Ú	AttentionÚ
fused_attné   TFç        NrT   Ú	num_headsÚqkv_biasÚqk_normÚ	attn_dropÚ	proj_dropÚ
norm_layerÚuse_rel_posÚ
input_sizeÚropec                 ó‚  •— ||dœ}t          ¦   «                              ¦   «          ||z  dk    s
J d¦   «         ‚|| _        ||z  | _        | j        dz  | _        t          ¦   «         | _        t          j        ||dz  fd|i|¤Ž| _	        |r || j        fi |¤Žnt          j
        ¦   «         | _        |r || j        fi |¤Žnt          j
        ¦   «         | _        t          j        |¦  «        | _        t          j        ||fi |¤Ž| _        t          j        |¦  «        | _        || _        | j        r†|
�J ‚|	€
J d¦   «         ‚t          j        t'          j        d|	d         z  d	z
  | j        fi |¤Ž¦  «        | _        t          j        t'          j        d|	d	         z  d	z
  | j        fi |¤Ž¦  «        | _        |
| _        d S )
N©Údevicer2   r   z$dim should be divisible by num_headsg      à¿é   ÚbiaszBInput size must be provided if using relative positional encoding.r-   r   )ÚsuperÚ__init__r_   Úhead_dimÚscaler   r\   ÚnnÚLinearÚqkvÚIdentityÚq_normÚk_normÚDropoutrb   Úprojrc   re   Ú	Parameterr;   ÚzerosrH   rI   rg   )ÚselfrT   r_   r`   ra   rb   rc   rd   re   rf   rg   rj   r2   ÚddÚ	__class__s                 €rD   rn   zAttention.__init__}   sØ  ø€ ð ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐØ�Y‰ !Ò#Ð#Ð#Ð%KÑ#Ô#Ð#Ø"ˆŒØ˜yÑ(ˆŒØ”] dÑ*ˆŒ
Ý(Ñ*Ô*ˆŒå”9˜S #¨¡'Ð?Ð?°Ð?¸BÐ?Ð?ˆŒØ9@ÐS�j�j ¤Ð5Ð5°"Ð5Ð5Ð5ÅbÄkÁmÄmˆŒØ9@ÐS�j�j ¤Ð5Ð5°"Ð5Ð5Ð5ÅbÄkÁmÄmˆŒÝœ IÑ.Ô.ˆŒÝ”I˜c 3Ð-Ð-¨"Ð-Ð-ˆŒ	Ýœ IÑ.Ô.ˆŒØ&ˆÔØÔð 	cØ�<�<�<àÐ&Ð&ØSñ 'Ô&Ð&õ  œ\­%¬+°a¸*ÀQ¼-Ñ6GÈ!Ñ6KÈTÌ]Ð*aÐ*aÐ^`Ð*aÐ*aÑbÔbˆDŒNÝœ\­%¬+°a¸*ÀQ¼-Ñ6GÈ!Ñ6KÈTÌ]Ð*aÐ*aÐ^`Ð*aÐ*aÑbÔbˆDŒNØˆŒ	ˆ	ˆ	rF   c                 ó  — |j         \  }}}}||z  }|                     ||d¦  «        }|                      |¦  «                             ||d| j        d¦  «                             ddddd¦  «        }|                     d|| j        z  |d¦  «                             d¦  «        \  }}	}
|                      |¦  «        |                      |	¦  «        }	}| j	        r"t          || j        | j        ||f||f¦  «        }nhd }| j        �_| j                             ¦   «         }t          ||¦  «                             |
¦  «        }t          |	|¦  «                             |
¦  «        }	| j        r=t$          j        j                             ||	|
|| j        r| j        j        nd¬¦  «        }nZ|| j        z  }||	                     d	d¦  «        z  }|�||z   }|                     d¬
¦  «        }|                      |¦  «        }||
z  }|                     || j        |d¦  «                             dd¦  «                             ||d¦  «        }|                      |¦  «        }|                      |¦  «        }|                     |||d¦  «        }|S )Nr.   rk   r-   r   r   é   r^   )Ú	attn_maskÚ	dropout_péþÿÿÿ)rT   )r6   r9   rs   Úviewr_   r:   Úunbindru   rv   re   rY   rH   rI   rg   Ú	get_embedr   Útype_asr\   r;   rq   Ú
functionalÚscaled_dot_product_attentionÚtrainingrb   Úprp   Ú	transposeÚsoftmaxrx   rc   )r{   ÚxrR   ÚHÚWrS   ÚNrs   rG   ÚkÚvrX   rg   Úattns                 rD   ÚforwardzAttention.forward¥   sm  € Ø”W‰
ˆˆ1ˆa�Ø�‰EˆØ�IŠI�a˜˜BÑÔˆØ�hŠh�q‰kŒk×Ò˜q ! Q¨¬¸Ñ;Ô;×CÒCÀAÀqÈ!ÈQÐPQÑRÔRˆà—+’+˜a  T¤^Ñ!3°Q¸Ñ;Ô;×BÒBÀ1ÑEÔE‰ˆˆ1ˆaà�{Š{˜1‰~Œ~˜tŸ{š{¨1™~œ~ˆ1ˆàÔð 	<Ý3°A°t´~ÀtÄ~ÐXYÐ[\ÐW]Ð`aÐcdÐ_eÑfÔfˆIˆIàˆIØŒyÐ$Ø”y×*Ò*Ñ,Ô,�Ý'¨¨4Ñ0Ô0×8Ò8¸Ñ;Ô;�Ý'¨¨4Ñ0Ô0×8Ò8¸Ñ;Ô;�àŒ?ð 	Ý”Ô#×@Ò@Ø�1�aØ#Ø.2¬mÐC˜$œ.Ô*Ð*Àð Añ ô ˆAˆAð �D”J‘ˆAØ�q—{’{ 2 rÑ*Ô*Ñ*ˆDØÐ$Ø˜iÑ'�Ø—<’< B�<Ñ'Ô'ˆDØ—>’> $Ñ'Ô'ˆDØ�q‘ˆAà�FŠF�1�d”n a¨Ñ,Ô,×6Ò6°q¸!Ñ<Ô<×DÒDÀQÈÈ2ÑNÔNˆØ�IŠI�a‰LŒLˆØ�NŠN˜1ÑÔˆØ�FŠF�1�a˜˜BÑÔˆØˆrF   )Ú__name__Ú
__module__Ú__qualname__r   ÚboolÚ__annotations__rq   Ú	LayerNormr4   Úfloatr   ÚModuler   r   rn   r”   Ú__classcell__©r}   s   @rD   r[   r[   z   s  ø€ € € € € € Ø�d”ÐÐÑð
 Ø!Ø!Ø!Ø!Ø*,¬,Ø %Ø48Ø(,ØØð&ð &àð&ð ð&ð ð	&ð
 ð&ð ð&ð ð&ð ˜RœYœð&ð ð&ð !  s¨C x¤Ô1ð&ð ˜2œ9Ô%ð&ð &ð &ð &ð &ð &ðP&ð &ð &ð &ð &ð &ð &rF   r[   c                   óâ   ‡ — e Zd Zdddddddej        ej        eddddddfdeded	ed
e	de	dedede
e         dedeej                 deej                 deej                 de	defˆ fd„Zd„ Zˆ xZS )ÚBlockç      @TFr^   Nr   rT   r_   Ú	mlp_ratior`   ra   rc   rb   Úinit_valuesÚ	drop_pathÚ	act_layerrd   Ú	mlp_layerre   Úwindow_sizec                 ó€  •— ||dœ}t          ¦   «                              ¦   «          || _         ||fi |¤Ž| _        t	          |f||||||||dk    r|n||f|dœ	|¤Ž| _        |rt          |fd|i|¤Žnt          j        ¦   «         | _	        |	dk    rt          |	¦  «        nt          j        ¦   «         | _         ||fi |¤Ž| _         |d|t          ||z  ¦  «        |
|dœ|¤Ž| _        |rt          |fd|i|¤Žnt          j        ¦   «         | _        |	dk    rt          |	¦  «        nt          j        ¦   «         | _        d S )Nri   r   )	r_   r`   ra   rb   rc   rd   re   rf   rg   r£   r^   )Úin_featuresÚhidden_featuresr¥   Údrop© )rm   rn   r§   Únorm1r[   r“   r   rq   rt   Úls1r   Ú
drop_path1Únorm2r4   ÚmlpÚls2Ú
drop_path2)r{   rT   r_   r¢   r`   ra   rc   rb   r£   r¤   r¥   rd   r¦   re   r§   rf   rg   rj   r2   r|   r}   s                       €rD   rn   zBlock.__init__Ð   s§  ø€ ð* ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐØ&ˆÔØ�Z Ð*Ð* rÐ*Ð*ˆŒ
ÝØð
àØØØØØ!Ø#Ø%0°AÒ%5Ð%5�z�z¸KÈÐ;UØð
ð 
ð ð
ð 
ˆŒ	ð FQÐc•:˜cÐAÐA¨{ÐA¸bÐAÐAÐAÕVXÔVaÑVcÔVcˆŒØ1:¸R²°�( 9Ñ-Ô-Ð-ÅRÄ[Á]Ä]ˆŒà�Z Ð*Ð* rÐ*Ð*ˆŒ
Ø�9ð 
ØÝ  i¡Ñ0Ô0ØØð	
ð 
ð
 ð
ð 
ˆŒð FQÐc•:˜cÐAÐA¨{ÐA¸bÐAÐAÐAÕVXÔVaÑVcÔVcˆŒØ1:¸R²°�( 9Ñ-Ô-Ð-ÅRÄ[Á]Ä]ˆŒˆˆrF   c           
      ód  — |j         \  }}}}|}|                      |¦  «        }d }| j        dk    rt          || j        ¦  «        \  }}|                      |                      |                      |¦  «        ¦  «        ¦  «        }| j        dk    rt          || j        ||f|¦  «        }||z   }|                     |||z  d¦  «        }||  	                    |  
                    |                      |                      |¦  «        ¦  «        ¦  «        ¦  «        z   }|                     |||d¦  «        }|S )Nr   r.   )r6   r­   r§   Úwindow_partitionr¯   r®   r“   Úwindow_unpartitionr9   r³   r²   r±   r°   )r{   r�   rR   rŽ   r�   rS   ÚshortcutÚpad_hws           rD   r”   zBlock.forward  s  € Ø”W‰
ˆˆ1ˆa�àˆØ�JŠJ�q‰MŒMˆà,0ˆØÔ˜aÒÐÝ(¨¨DÔ,<Ñ=Ô=‰IˆAˆvà�OŠO˜DŸHšH T§Y¢Y¨q¡\¤\Ñ2Ô2Ñ3Ô3ˆð Ô˜aÒÐÝ" 1 dÔ&6¸¸A¸ÀÑGÔGˆAà�q‰Lˆà�IŠI�a˜˜Q™ Ñ#Ô#ˆØ�—’ §¢¨¯ª°$·*²*¸Q±-´-Ñ)@Ô)@Ñ AÔ AÑBÔBÑBˆØ�IŠI�a˜˜A˜rÑ"Ô"ˆàˆrF   )r•   r–   r—   rq   ÚGELUrš   r   r4   r›   r˜   r   r   rœ   rn   r”   r�   rž   s   @rD   r    r    Î   s=  ø€ € € € € ð  "Ø!Ø!Ø!Ø!Ø+/Ø!Ø)+¬Ø*,¬,Ø),Ø %Ø ØØØØð'2Sð 2Sàð2Sð ð2Sð ð	2Sð
 ð2Sð ð2Sð ð2Sð ð2Sð " %œð2Sð ð2Sð ˜BœI”ð2Sð ˜RœYœð2Sð ˜BœI”ð2Sð ð2Sð ð2Sð 2Sð 2Sð 2Sð 2Sð 2Sðhð ð ð ð ð ð rF   r    r�   r§   c           	      ó`  — | j         \  }}}}|||z  z
  |z  }|||z  z
  |z  }t          j        | ddd|d|f¦  «        } ||z   ||z   }	}|                      |||z  ||	|z  ||¦  «        } |                      dddddd¦  «                             ¦   «                              d|||¦  «        }
|
||	ffS )aU  
    Partition into non-overlapping windows with padding if needed.
    Args:
        x (tensor): input tokens with [B, H, W, C].
        window_size (int): window size.

    Returns:
        windows: windows after partition with [B * num_windows, window_size, window_size, C].
        (Hp, Wp): padded height and width before partition
    r   r   rk   r-   r   é   r.   )r6   r7   Úpadrƒ   r:   Ú
contiguous)r�   r§   rR   rŽ   r�   ÚCÚpad_hÚpad_wÚHpÚWpÚwindowss              rD   rµ   rµ     sÙ   € ð ”�J€A€qˆ!ˆQà˜1˜{™?Ñ*¨kÑ9€EØ˜1˜{™?Ñ*¨kÑ9€EÝ	Œˆa�!�Q˜˜5 ! UÐ+Ñ,Ô,€AØ�‰Y˜˜E™	ˆ€Bà	�Šˆq�"˜Ñ# [°"¸Ñ2CÀ[ÐRSÑTÔT€AØ�iŠi˜˜1˜a  A qÑ)Ô)×4Ò4Ñ6Ô6×;Ò;¸BÀÈ[ÐZ[Ñ\Ô\€GØ�R˜�HÐÐrF   rÃ   Úhwr¸   c                 ód  — |�|n|\  }}|\  }}| j         d         ||z  |z  |z  z  }|                      |||z  ||z  ||d¦  «        }	|	                     dddddd¦  «                             ¦   «                              |||d¦  «        }	|	dd…d|…d|…dd…f                              ¦   «         }	|	S )	a›  
    Window unpartition into original sequences and removing padding.
    Args:
        windows (tensor): input tokens with [B * num_windows, window_size, window_size, C].
        window_size (int): window size.
        pad_hw (Tuple): padded height and width (Hp, Wp).
        hw (Tuple): original height and width (H, W) before padding.

    Returns:
        x: unpartitioned sequences with [B, H, W, C].
    Nr   r.   r   rk   r-   r   r»   )r6   rƒ   r:   r½   )
rÃ   r§   rÄ   r¸   rÁ   rÂ   rŽ   r�   rR   r�   s
             rD   r¶   r¶   4  sÑ   € ð Ð)ˆVˆV¨r�F€BˆØ�D€A€qØŒ�aÔ˜R "™W¨Ñ3°{ÑBÑC€AØ�Š�Q˜˜kÑ)¨2°Ñ+<¸kÈ;ÐXZÑ[Ô[€AØ	�	Š	�!�Q˜˜1˜a Ñ#Ô#×.Ò.Ñ0Ô0×5Ò5°a¸¸RÀÑDÔD€AØ	ˆ!ˆ!ˆ!ˆRˆaˆR��!��Q�Q�Qˆ,Œ×"Ò"Ñ$Ô$€AØ€HrF   c            H       ó   ‡ — e Zd ZdZdddddddddd	d
d	ddddddd eeej        d	¬¦  «        ej	        ej
        eedd	d	ddddd
d
d
d
f#dededededededededededee         dededed ed!ed"ed#ed$ed%eej                 d&eeej                          d'eeej                          d(eej                 d)eej                 d*ed+ed,ed-ed.eed/f         d0ed1ed2ee         d3eeeeef         eeef         f                  fBˆ fd4„Zej        j        d5„ ¦   «         Zej        j        dJd6„¦   «         Zej        j        dKd7„¦   «         Zej        j        d8ej        fd9„¦   «         ZdLded1ee         fd:„Z	 	 	 	 	 dMd<ej        d=eeee e         f                  d>ed?ed@edAed8ee ej                 eej        e ej                 f         f         fdB„Z!	 	 	 dNd=eeee e         f                  dCedDefdE„Z"dF„ Z#dJdGefdH„Z$dI„ Z%ˆ xZ&S )Or'   zâ Vision Transformer for Segment-Anything Model(SAM)

    A PyTorch impl of : `Exploring Plain Vision Transformer Backbones for Object Detection` or `Segment Anything Model (SAM)`
        - https://arxiv.org/abs/2010.11929
    é   é   rk   é   é   r¡   TFNr^   Ú )Ú
output_fmtÚstrict_img_sizeé   r¬   é   ÚavgÚimg_sizeÚ
patch_sizeÚin_chansÚnum_classesÚ	embed_dimÚdepthr_   r¢   r`   ra   r£   Úpre_normÚ	drop_rateÚpos_drop_rateÚpatch_drop_rateÚproj_drop_rateÚattn_drop_rateÚdrop_path_rateÚweight_initÚembed_layerrd   r¥   Úblock_fnr¦   Úuse_abs_posre   Úuse_roper§   Úglobal_attn_indexes.Ú
neck_chansÚglobal_poolÚhead_hidden_sizeÚref_feat_shapec$                 óî  •‡ ‡‡‡‡	‡
‡‡‡‡‡‡‡‡‡‡‡&‡'‡(‡)— t          ¦   «                              ¦   «          |"|#dœŠ&‰pt          t          j        d¬¦  «        Š‰pt          j        Š|‰ _        |‰ _        |‰ _        ‰x‰ _	        x‰ _
        ‰ _        d‰ _         |d|||‰| dœ‰&¤Ž‰ _        ‰ j        j        Š(t          ‰ j        d¦  «        r‰ j                             ¦   «         n|Š)|r9t          j        t%          j        d‰(d         ‰(d         ‰fi ‰&¤Ž¦  «        ‰ _        nd	‰ _        t          j        |¬
¦  «        ‰ _        |dk    rt/          |d¬¦  «        ‰ _        nt          j        ¦   «         ‰ _        |r	 ‰‰fi ‰&¤Žnt          j        ¦   «         ‰ _        |r–‰r
J d¦   «         ‚|!�@t7          |!¦  «        dk    sJ ‚t9          |!d         ¦  «        }$t9          |!d         ¦  «        }%nd	x}$}%t;          ‰‰z  d‰(|$¬¦  «        ‰ _        t;          ‰‰z  dt9          ‰¦  «        |%¬¦  «        ‰ _        nd	‰ _        d	‰ _        tA          ||¦  «        Š't          j!        ˆˆˆˆ&ˆ'ˆˆˆ(ˆˆˆˆˆˆˆ
ˆ	ˆ ˆˆfd„tE          |¦  «        D ¦   «         Ž ‰ _#        ˆˆ)fd„tE          |¦  «        D ¦   «         ‰ _$        |rct          j!        t          j%        ‰|fdddœ‰&¤ŽtM          |fi ‰&¤Žt          j%        ||fddddœ‰&¤ŽtM          |fi ‰&¤Ž¦  «        ‰ _'        |‰ _	        n/| rt          j        ¦   «         ‰ _'        ntM          ‰fi ‰&¤Ž‰ _'        ‰}| rtQ          ||f| ||dœ‰&¤Ž‰ _)        d	S tU          ||f||dœ‰&¤Ž‰ _)        d	S )a×  
        Args:
            img_size: Input image size.
            patch_size: Patch size.
            in_chans: Number of image input channels.
            num_classes: Number of classes for classification head.
            global_pool: Type of global pooling for final sequence (default: 'token').
            embed_dim: Transformer embedding dimension.
            depth: Depth of transformer.
            num_heads: Number of attention heads.
            mlp_ratio: Ratio of mlp hidden dim to embedding dim.
            qkv_bias: Enable bias for qkv projections if True.
            init_values: Layer-scale init values (layer-scale enabled if not None).
            drop_rate: Head dropout rate.
            pos_drop_rate: Position embedding dropout rate.
            attn_drop_rate: Attention dropout rate.
            drop_path_rate: Stochastic depth rate.
            weight_init: Weight initialization scheme.
            embed_layer: Patch embedding layer.
            norm_layer: Normalization layer.
            act_layer: MLP activation layer.
            block_fn: Transformer block layer.
            use_abs_pos: If True, use absolute positional embeddings.
            use_rel_pos: If True, add relative positional embeddings to the attention map.
            use_rope: If True, add rotary position embeddings to q/k in attention block.
            window_size: Window size for window attention blocks. If 0, not use window attention.
            global_attn_indexes: Indexes for blocks using global attention. Used when window_size > 0.
            global_pool: Global pooling type.
            head_hidden_size: If set, use NormMlpHead
            ref_feat_shape: Tuple of reference feature shapes for ROPE, (global, local)
        ri   g�íµ ÷Æ°>)ÚepsF)rÑ   rÒ   rÓ   rÕ   rl   Ú
feat_ratior   r   N)rŠ   )Únum_prefix_tokenszCROPE and relative pos embeddings should not be enabled at same timer-   )Ú	in_pixelsÚ
feat_shaperç   c           	      ó¸   •— g | ]V} ‰di d ‰“d‰“d‰“d‰“d‰“d‰
“d‰“d‰“d‰|         “d	‰“d
‰“d‰“d‰“d|‰vr‰nd“d‰	“d|‰vr‰j         n‰j        “‰¤Ž‘ŒWS )rT   r_   r¢   r`   ra   r£   rc   rb   r¤   rd   r¥   r¦   re   r§   r   rf   rg   r¬   )Úrope_windowÚrope_global)Ú.0Úir¥   rÜ   rà   r|   ÚdprrÕ   rã   Ú	grid_sizer£   r¦   r¢   rd   r_   rÛ   ra   r`   r{   re   r§   s     €€€€€€€€€€€€€€€€€€€rD   ú
<listcomp>z1VisionTransformerSAM.__init__.<locals>.<listcomp>Ö  s'  ø€ ð &#ð &#ð &#ð( ð' ˆHð ð ð Ø�Iðà#˜)ðð $˜)ðð "˜ð	ð
  ˜ðð (˜Kðð )˜.ðð )˜.ðð ˜aœ&˜&ðð &˜:ðð $˜)ðð $˜)ðð (˜Kðð ,-Ð4GÐ+GÐ+G˜K˜KÈQøðð %˜9ðð  *+Ð2EÐ)EÐ)E�TÔ%Ð%È4ÔK[øØð#ð ð&#ð &#ð &#rF   c                 ó:   •— g | ]}t          d |› �‰‰¬¦  «        ‘ŒS )zblocks.)ÚmoduleÚnum_chsÚ	reduction©Údict)rñ   rò   rÕ   Úrs     €€rD   rõ   z1VisionTransformerSAM.__init__.<locals>.<listcomp>ë  sA   ø€ ð ^ð ^ð ^ØKL�D˜ !˜˜¨yÀAÐFÑFÔFð^ð ^ð ^rF   )Úkernel_sizerl   rk   )rý   Úpaddingrl   )Úhidden_sizeÚ	pool_typerØ   )r   rØ   r¬   )+rm   rn   r   rq   rš   r¹   rÔ   rÓ   rå   Únum_featuresræ   rÕ   Úgrad_checkpointingÚpatch_embedrô   Úhasattrrê   ry   r;   rz   Ú	pos_embedrw   Úpos_dropr   Ú
patch_droprt   Únorm_preÚlenr   r   rð   rï   r   Ú
SequentialÚrangeÚblocksÚfeature_infoÚConv2dr   Úneckr   Úheadr   )+r{   rÑ   rÒ   rÓ   rÔ   rÕ   rÖ   r_   r¢   r`   ra   r£   r×   rØ   rÙ   rÚ   rÛ   rÜ   rÝ   rÞ   rß   rd   r¥   rà   r¦   rá   re   râ   r§   rã   rä   rå   ræ   rç   rj   r2   Úref_feat_shape_globalÚref_feat_shape_windowr|   ró   rô   rü   r}   s+   `    ` `````    ``   ```` ` ``        @@@@€rD   rn   zVisionTransformerSAM.__init__R  sö  øøøøøøøøøøøøøøøøøøøøø€ õJ 	‰Œ×ÒÑÔÐØ¨Ð/Ð/ˆØÐB¥7­2¬<¸TÐ#BÑ#BÔ#Bˆ
ØÐ(¥¤ˆ	à&ˆÔØ ˆŒØ&ˆÔØENÐNˆÔÐN˜DÔ1°D´NØ"'ˆÔà&˜;ð 
ØØ!ØØØ�ð
ð 
ð ð
ð 
ˆÔð Ô$Ô.ˆ	Ý-4°TÔ5EÀ|Ñ-TÔ-TÐdˆDÔ×'Ò'Ñ)Ô)Ð)ÐZdˆàð 	"åœ\­%¬+°a¸À1¼ÀyÐQRÄ|ÐU^Ð*eÐ*eÐbdÐ*eÐ*eÑfÔfˆDŒNˆNà!ˆDŒNÝœ
 ]Ð3Ñ3Ô3ˆŒØ˜QÒÐÝ*ØØ"#ðñ ô ˆDŒOˆOõ
 !œk™mœmˆDŒOØ7?ÐR˜
˜
 9Ð3Ð3°Ð3Ð3Ð3ÅRÄ[Á]Ä]ˆŒàð 	$Ø"ÐiÐiÐ$iÑiÔi�?ØÐ)Ý˜>Ñ*Ô*¨aÒ/Ð/Ð/Ð/Ý(1°.ÀÔ2CÑ(DÔ(DÐ%Ý(1°.ÀÔ2CÑ(DÔ(DÐ%Ð%à@DÐDÐ%Ð(=Ý1Ø˜YÑ&ØØ$Ø4ð	 ñ  ô  ˆDÔõ  2Ø˜YÑ&ØÝ$ [Ñ1Ô1Ø4ð	 ñ  ô  ˆDÔÐð  $ˆDÔØ#ˆDÔõ (¨¸Ñ>Ô>ˆÝ”mð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#ð &#õ( ˜5‘\”\ð)&#ñ &#ô &#ð $ˆŒð*^ð ^ð ^ð ^ð ^ÝPUÐV[ÑP\ÔP\ð^ñ ^ô ^ˆÔð ð 	#ÝœÝ”	ØØðð !"Øð	ð ð
 ðð õ ˜JÐ-Ð-¨"Ð-Ð-Ý”	ØØðð !"ØØðð ð ðð õ ˜JÐ-Ð-¨"Ð-Ð-ñ#ô ˆDŒIð& !+ˆDÔÐàð 9ÝœK™MœM�”	�	õ (¨	Ð8Ð8°RÐ8Ð8�”	Ø"ˆJð ð 	Ý-ØØðð -Ø%Ø#ðð ð ðð ˆDŒIˆIˆIõ 'ØØðð &Ø#ð	ð ð
 ðð ˆDŒIˆIˆIrF   c                 ó
   — ddhS )Nr  Ú
dist_tokenr¬   ©r{   s    rD   Úno_weight_decayz$VisionTransformerSAM.no_weight_decay  s   € à˜\Ð*Ð*rF   c                 ó(   — t          dddg¬¦  «        S )Nz^pos_embed|patch_embed)z^blocks\.(\d+)N)z^norm)iŸ† )Ústemr  rú   )r{   Úcoarses     rD   Úgroup_matcherz"VisionTransformerSAM.group_matcher"  s%   € åØ*Ø-Ð/CÐDð
ñ 
ô 
ð 	
rF   c                 ó   — || _         d S ©N)r  )r{   Úenables     rD   Úset_grad_checkpointingz+VisionTransformerSAM.set_grad_checkpointing)  s   € à"(ˆÔÐÐrF   r+   c                 ó   — | j         S r  ©r  r  s    rD   Úget_classifierz#VisionTransformerSAM.get_classifier-  s
   € àŒyÐrF   c                 óZ   — t          | ¦  «        }|| _         | j        j        ||fi |¤Ž d S r  )r   rÔ   r  Úreset)r{   rÔ   rå   r|   s       rD   Úreset_classifierz%VisionTransformerSAM.reset_classifier1  s;   € Ý˜dÑ#Ô#ˆØ&ˆÔØˆŒ	Œ˜ [Ð7Ð7°BÐ7Ð7Ð7Ð7Ð7rF   ÚNCHWr�   ÚindicesÚnormÚ
stop_earlyrÌ   Úintermediates_onlyc                 óÖ  — |dk    s
J d¦   «         ‚g }t          t          | j        ¦  «        |¦  «        \  }}	|                      |¦  «        }| j        �%|t          | j        |j        dd…         ¦  «        z   }|                      |¦  «        }|                      |¦  «        }|  	                    |¦  «        }t          j                             ¦   «         s|s| j        }
n| j        d|	dz   …         }
t          |
¦  «        D ]¶\  }}| j        r/t          j                             ¦   «         st          ||¦  «        }n ||¦  «        }||v rl|r?|                     |                      |                     dddd¦  «        ¦  «        ¦  «         Œ‹|                     |                     dddd¦  «        ¦  «         Œ·|r|S |                      |                     dddd¦  «        ¦  «        }||fS )aè   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 all 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:

        r%  z&Output shape for ViT-SAM must be NCHW.Nr   rk   r   r-   )r!   r	  r  r  r  r   r6   r  r  r  r;   ÚjitÚis_scriptingÚ	enumerater  r#   Úappendr  r:   )r{   r�   r&  r'  r(  rÌ   r)  ÚintermediatesÚtake_indicesÚ	max_indexr  rò   Úblks                rD   Úforward_intermediatesz*VisionTransformerSAM.forward_intermediates6  sõ  € ð* ˜VÒ#Ð#Ð#Ð%MÑ#Ô#Ð#ØˆÝ"6µs¸4¼;Ñ7GÔ7GÈÑ"QÔ"QÑˆ�ið ×Ò˜QÑÔˆØŒ>Ð%àÕ/°´ÀÄÈÈ!ÈÄÑMÔMÑMˆAØ�MŠM˜!ÑÔˆØ�OŠO˜AÑÔˆØ�MŠM˜!ÑÔˆåŒ9×!Ò!Ñ#Ô#ð 	1¨:ð 	1Ø”[ˆFˆFà”[  )¨a¡- Ô0ˆFÝ Ñ'Ô'ð 	@ð 	@‰FˆAˆsØÔ&ð ­u¬y×/EÒ/EÑ/GÔ/Gð Ý˜s AÑ&Ô&��à�C˜‘F”F�Ø�LÐ Ð àð @ð "×(Ò(¨¯ª°1·9²9¸QÀÀ1ÀaÑ3HÔ3HÑ)IÔ)IÑJÔJÐJÐJà!×(Ò(¨¯ª°1°a¸¸AÑ)>Ô)>Ñ?Ô?Ð?øàð 	!Ø Ð à�IŠI�a—i’i  1 a¨Ñ+Ô+Ñ,Ô,ˆà�-ÐÐrF   Ú
prune_normÚ
prune_headc                 óâ   — t          t          | j        ¦  «        |¦  «        \  }}| j        d|dz   …         | _        |rt          j        ¦   «         | _        |r|                      dd¦  «         |S )z@ Prune layers not required for specified intermediates.
        Nr   r   rË   )r!   r	  r  rq   rt   r  r$  )r{   r&  r4  r5  r0  r1  s         rD   Úprune_intermediate_layersz.VisionTransformerSAM.prune_intermediate_layersq  sq   € õ #7µs¸4¼;Ñ7GÔ7GÈÑ"QÔ"QÑˆ�iØ”k . 9¨q¡= .Ô1ˆŒØð 	&åœ™œˆDŒIØð 	)Ø×!Ò! ! RÑ(Ô(Ð(ØÐrF   c                 óü  — |                       |¦  «        }| j        �%|t          | j        |j        dd…         ¦  «        z   }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }| j        r4t          j	         
                    ¦   «         st          | j        |¦  «        }n|                      |¦  «        }|                      |                     dddd¦  «        ¦  «        }|S )Nr   rk   r   r-   )r  r  r   r6   r  r  r  r  r;   r+  r,  r$   r  r  r:   ©r{   r�   s     rD   Úforward_featuresz%VisionTransformerSAM.forward_features‚  sÛ   € Ø×Ò˜QÑÔˆØŒ>Ð%àÕ/°´ÀÄÈÈ!ÈÄÑMÔMÑMˆAØ�MŠM˜!ÑÔˆØ�OŠO˜AÑÔˆØ�MŠM˜!ÑÔˆØÔ"ð 	­5¬9×+AÒ+AÑ+CÔ+Cð 	Ý˜tœ{¨AÑ.Ô.ˆAˆAà—’˜A‘”ˆAØ�IŠI�a—i’i  1 a¨Ñ+Ô+Ñ,Ô,ˆØˆrF   Ú
pre_logitsc                 ó^   — |r|                       |d¬¦  «        n|                       |¦  «        S )NT)r;  r   )r{   r�   r;  s      rD   Úforward_headz!VisionTransformerSAM.forward_head‘  s,   € Ø0:ÐLˆt�yŠy˜ tˆyÑ,Ô,Ð,ÀÇ	Â	È!ÁÄÐLrF   c                 óZ   — |                       |¦  «        }|                      |¦  «        }|S r  )r:  r=  r9  s     rD   r”   zVisionTransformerSAM.forward”  s-   € Ø×!Ò! !Ñ$Ô$ˆØ×Ò˜aÑ Ô ˆØˆrF   ©F)Tr  )NFFr%  F)NFT)'r•   r–   r—   Ú__doc__r   r   r   ÚNHWCrq   rš   r¹   r    r   r4   r›   r˜   r   Ústrr   rœ   r   rn   r;   r+  Úignorer  r  r  r!  r$  ÚTensorr	   r   r3  r7  r:  r=  r”   r�   rž   s   @rD   r'   r'   K  s¢  ø€ € € € € ðð ð !Ø ØØ"Ø ØØØ!Ø!Ø!Ø+/Ø"Ø!Ø#%Ø%'Ø$&Ø$&Ø$&Ø!Ø+2¨7°:È&Ì+ÐglÐ+mÑ+mÔ+mØ46´LØ35´7Ø(-Ø),Ø $Ø %Ø"Ø!Ø35Ø!Ø$Ø.2ØPTØØðIJð JàðJð ðJð ð	Jð
 ðJð ðJð ðJð ðJð ðJð ðJð ðJð " %œðJð ðJð ðJð !ðJð  #ð!Jð" "ð#Jð$ "ð%Jð& "ð'Jð( ð)Jð* ˜bœiœð+Jð, !  b¤i¤Ô1ð-Jð.    R¤Y¤Ô0ð/Jð0 ˜2œ9”oð1Jð2 ˜BœI”ð3Jð4 ð5Jð6 ð7Jð8 ð9Jð: ð;Jð< "' s¨C x¤ð=Jð> ð?Jð@ ðAJðB ' sœmðCJðD % U¨5°°c°¬?¸EÀ#ÀsÀ(¼OÐ+KÔ%LÔMðEJð Jð Jð Jð Jð JðX „YÔð+ð +ñ Ôð+ð „YÔð
ð 
ð 
ñ Ôð
ð „YÔð)ð )ð )ñ Ôð)ð „YÔð ¤	ð ð ð ñ Ôðð8ð 8¨Cð 8¸hÀs¼mð 8ð 8ð 8ð 8ð 8<ØØ$Ø$Ø',ð9 ð 9 àŒ|ð9 ð ˜e C¨¨c¬ NÔ3Ô4ð9 ð ð	9 ð
 ð9 ð ð9 ð !%ð9 ð 
ˆt�E”LÔ! 5¨¬°t¸E¼LÔ7IÐ)IÔ#JÐJÔ	Kð9 ð 9 ð 9 ð 9 ðz 8<Ø$Ø#ð	ð à˜e C¨¨c¬ NÔ3Ô4ðð ðð ð	ð ð ð ð"ð ð ðMð M¨$ð Mð Mð Mð Mðð ð ð ð ð ð rF   c                 óÂ   — d| v }i }|                       ¦   «         D ]C\  }}|                     d¦  «        r!|dd…         }|                     dd¦  «        }n|rŒ>|||<   ŒD|S )z Remap SAM checkpoints -> timm z%image_encoder.patch_embed.proj.weightzimage_encoder.rÎ   Nzmlp.linzmlp.fc)ÚitemsÚ
startswithÚreplace)Ú
state_dictÚmodelÚsam_checkpointÚout_dictr‘   r’   s         rD   Úcheckpoint_filter_fnrM  š  s„   € ð
 =À
ÐJ€NØ€HØ× Ò Ñ"Ô"ð ð ‰ˆˆ1Ø�<Š<Ð(Ñ)Ô)ð 	Ø�"�#�#”ˆAØ—	’	˜) XÑ.Ô.ˆAˆAàð ØØˆ�‰ˆØ€OrF   rË   c                 ó6   — | ddd dddt           t          dddœ|¥S )	Néè  ©rk   rÇ   rÇ   çÍÌÌÌÌÌì?ÚbicubicTzpatch_embed.projzhead.fc)ÚurlrÔ   rf   Ú	pool_sizeÚcrop_pctÚinterpolationÚfixed_input_sizeÚmeanÚstdÚ
first_convÚ
classifier)r   r   )rS  Úkwargss     rD   Ú_cfgr]  ¬  s7   € àØ¨?ÈØ¨ÈÝ'Õ0FØ(¸	ðð ð ðð rF   zDhttps://dl.fbaipublicfiles.com/segment_anything/sam_vit_b_01ec64.pthztimm/z
apache-2.0rP  r3   )rS  Ú	hf_hub_idÚlicenserX  rY  rÔ   rf   rU  zDhttps://dl.fbaipublicfiles.com/segment_anything/sam_vit_l_0b3195.pthzDhttps://dl.fbaipublicfiles.com/segment_anything/sam_vit_h_4b8939.pthrO  )rk   éà   r`  rQ  )rX  rY  rÔ   rf   rU  )zsamvit_base_patch16.sa1bzsamvit_large_patch16.sa1bzsamvit_huge_patch16.sa1bÚsamvit_base_patch16_224Fc                 ó„   — |                      dd¦  «        }t          t          | |ft          t	          |d¬¦  «        dœ|¤ŽS )NÚout_indicesrk   Úgetter)rc  Úfeature_cls)Úpretrained_filter_fnÚfeature_cfg)Úpopr    r'   rM  rû   )ÚvariantÚ
pretrainedr\  rc  s       rD   Ú_create_vision_transformerrk  Ó  sY   € Ø—*’*˜]¨AÑ.Ô.€KÝÝØØðõ 2Ý [¸hÐGÑGÔGðð ð ðð ð rF   c           
      ól   — t          ddddg d¢ddd¬¦  «        }t          	 dd
| it          |fi |¤Ž¤Ž}|S )z# ViT-B/16 for Segment-Anything
    rÈ   rÉ   rÊ   ©r-   r»   r]   é   rÎ   TrÇ   ©rÒ   rÕ   rÖ   r_   rã   r§   re   rÑ   Úsamvit_base_patch16rj  )rp  ©rû   rk  ©rj  r\  Ú
model_argsrJ  s       rD   rp  rp  ß  st   € õ Ø ¨B¸"ÐR_ÐR_ÐR_Ø D°4ðñ ô €Jõ 'ØðTð TØ*4ðTÝ8<¸ZÐ8RÐ8RÈ6Ð8RÐ8RðTð T€Eà€LrF   c           
      ól   — t          ddddg d¢ddd¬¦  «        }t          	 d
d	| it          |fi |¤Ž¤Ž}|S )z# ViT-L/16 for Segment-Anything
    rÈ   rÇ   é   )r»   rn  é   é   rÎ   Tro  Úsamvit_large_patch16rj  )rx  rq  rr  s       rD   rx  rx  ì  st   € õ Ø ¨R¸2ÐSbÐSbÐSbØ D°4ðñ ô €Jõ 'ØðUð UØ+5ðUÝ9=¸jÐ9SÐ9SÈFÐ9SÐ9SðUð U€Eà€LrF   c           
      ól   — t          ddddg d¢ddd¬¦  «        }t          	 dd
| it          |fi |¤Ž¤Ž}|S )z# ViT-H/16 for Segment-Anything
    rÈ   i   é    )é   é   rw  é   rÎ   TrÇ   ro  Úsamvit_huge_patch16rj  )r~  rq  rr  s       rD   r~  r~  ù  st   € õ Ø ¨R¸2ÐSbÐSbÐSbØ D°4ðñ ô €Jõ 'ØðTð TØ*4ðTÝ8<¸ZÐ8RÐ8RÈ6Ð8RÐ8RðTð T€Eà€LrF   c                 óp   — t          ddddg d¢ddddd	¬
¦
  «
        }t          	 dd| it          |fi |¤Ž¤Ž}|S )z# ViT-B/16 based on samvit arch
    rÈ   rÉ   rÊ   rm  rÎ   TFr`  N)
rÒ   rÕ   rÖ   r_   rã   r§   re   rá   rÑ   rä   ra  rj  )ra  rq  rr  s       rD   ra  ra    sy   € õ Ø ¨B¸"ÐR_ÐR_ÐR_Ø D°eÀcÐVZðñ ô €Jõ 'Ø!ðXð XØ.8ðXÝ<@ÀÐ<VÐ<VÈvÐ<VÐ<VðXð X€Eà€LrF   r  )rË   r?  )Kr@  ÚloggingÚ	functoolsr   Útypingr   r   r   r   r   r	   r;   Útorch.nnrq   Útorch.nn.functionalr‡   r7   Ú	timm.datar
   r   r   r   Útimm.layersr   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   Ú	torch.jitr   Ú_builderr    Ú	_featuresr!   Ú_features_fxr"   Ú_manipulater#   r$   Ú	_registryr%   r&   Ú__all__Ú	getLoggerr•   Ú_loggerr4   rD  rE   rY   rœ   r[   r    rµ   r¶   r'   rM  r]  Údefault_cfgsrk  rp  rx  r~  ra  r¬   rF   rD   ú<module>r‘     s°  ðð
ð 
ð €€€Ø Ð Ð Ð Ð Ð Ø ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?à €€€Ø Ð Ð Ð Ð Ð Ø Ð Ð Ð Ð Ð Ð Ð Ð Ø rÐ rÐ rÐ rÐ rÐ rÐ rÐ rÐ rÐ rÐ rÐ rðð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð$ Ð Ð Ð Ð Ð à *Ð *Ð *Ð *Ð *Ð *Ø +Ð +Ð +Ð +Ð +Ð +Ø 3Ð 3Ð 3Ð 3Ð 3Ð 3Ø 3Ð 3Ð 3Ð 3Ð 3Ð 3Ð 3Ð 3Ø <Ð <Ð <Ð <Ð <Ð <Ð <Ð <ð "Ð
"€ð ˆ'Ô
˜HÑ
%Ô
%€ð3˜ð 3 Sð 3°5´<ð 3ÀEÄLð 3ð 3ð 3ð 3ð@ Ð ˜+Ñ &Ô &Ð &ð7ØŒ<ð7à”<ð7ð ”<ð7ð �c˜3�h”ð	7ð
 �c˜3�h”ð7ð „\ð7ð 7ð 7ð 7ðDQð Qð Qð Qð Q�”	ñ Qô Qð QðhLð Lð Lð Lð LˆBŒIñ Lô Lð Lð^˜œð °3ð ¸5ÀÄÈuÐUXÐZ]ÐU]ÌÐA^Ô;_ð ð ð ð ð0 gkðð ØŒ\ðØ(+ðØ16°s¸C°x´ðØJRÐSXÐY\Ð^aÐYaÔSbÔJcðà
„\ðð ð ð ð.Lð Lð Lð Lð L˜2œ9ñ Lô Lð Lð^
ð ð ð$ð ð ð ð %Ð$ð !% ØRØØØ"Ð(<È!Ø"¨Sð!2ñ !2ô !2ð "& ØRØØØ"Ð(<È!Ø"¨Sð"2ñ "2ô "2ð !% ØRØØØ"Ð(<È!Ø"¨Sð!2ñ !2ô !2ð  $˜tØ"Ð(<È$Ø ¨3ð 0ñ  0ô  0ð-&ð &ñ ô €ð8	ð 	ð 	ð 	ð ð	ð 	Ð7Kð 	ð 	ð 	ñ „ð	ð ð	ð 	Ð8Lð 	ð 	ð 	ñ „ð	ð ð	ð 	Ð7Kð 	ð 	ð 	ñ „ð	ð ð	ð 	Ð;Oð 	ð 	ð 	ñ „ð	ð 	ð 	rF   