§
    ·¯¸jBO  ã                   óŒ  — d Z ddlZddlmZ ddl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 ddlmZ dd	lmZmZ d
gZd„ Z G d„ de
j        ¦  «        Z G d„ de
j        ¦  «        Z G d„ de¦  «        Z G d„ de¦  «        Z G d„ de¦  «        Z  G d„ de
j        ¦  «        Z! G d„ d
e
j        ¦  «        Z"d4d„Z#d5d„Z$ e e$d¬¦  «         e$dd ¬!¦  «         e$d"¬¦  «         e$d#¬¦  «         e$d$¬¦  «         e$d%¬¦  «         e$d&d ¬!¦  «         e$d'¬¦  «         e$d(¬¦  «        d)œ	¦  «        Z%ed4d*e"fd+„¦   «         Z&ed4d*e"fd,„¦   «         Z'ed4d*e"fd-„¦   «         Z(ed4d*e"fd.„¦   «         Z)ed4d*e"fd/„¦   «         Z*ed4d*e"fd0„¦   «         Z+ed4d*e"fd1„¦   «         Z,ed4d*e"fd2„¦   «         Z-ed4d*e"fd3„¦   «         Z.dS )6a:  
SEResNet implementation from Cadene's pretrained models
https://github.com/Cadene/pretrained-models.pytorch/blob/master/pretrainedmodels/models/senet.py
Additional credit to https://github.com/creafz

Original model: https://github.com/hujie-frank/SENet

ResNet code gently borrowed from
https://github.com/pytorch/vision/blob/master/torchvision/models/resnet.py

FIXME I'm deprecating this model and moving them to ResNet as I don't want to maintain duplicate
support for extras like dilation, switchable BN/activations, feature extraction, etc that don't exist here.
é    N)ÚOrderedDict)ÚTypeÚOptionalÚTuple©ÚIMAGENET_DEFAULT_MEANÚIMAGENET_DEFAULT_STD)Úcreate_classifierÚget_device_dtypeé   )Úbuild_model_with_cfg)Úregister_modelÚgenerate_default_cfgsÚSENetc                 óX  — t          | t          j        ¦  «        r)t          j                             | j        dd¬¦  «         d S t          | t          j        ¦  «        rLt          j                             | j        d¦  «         t          j                             | j        d¦  «         d S d S )NÚfan_outÚrelu)ÚmodeÚnonlinearityg      ð?ç        )	Ú
isinstanceÚnnÚConv2dÚinitÚkaiming_normal_ÚweightÚBatchNorm2dÚ	constant_Úbias)Úms    úa/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/timm/models/senet.pyÚ_weight_initr"      s“   € Ý�!•R”YÑÔð &Ý
Œ×Ò ¤¨yÀvÐÑNÔNÐNÐNÐNÝ	�A•r”~Ñ	&Ô	&ð &Ý
Œ×Ò˜!œ( BÑ'Ô'Ð'Ý
Œ×Ò˜!œ& "Ñ%Ô%Ð%Ð%Ð%ð&ð &ó    c                   ó0   ‡ — e Zd Zddedefˆ fd„Zd„ Zˆ xZS )ÚSEModuleNÚchannelsÚ	reductionc                 ó*  •— ||dœ}t          ¦   «                              ¦   «          t          j        |||z  fddi|¤Ž| _        t          j        d¬¦  «        | _        t          j        ||z  |fddi|¤Ž| _        t          j        ¦   «         | _	        d S )N©ÚdeviceÚdtypeÚkernel_sizer   T©Úinplace)
ÚsuperÚ__init__r   r   Úfc1ÚReLUr   Úfc2ÚSigmoidÚsigmoid)Úselfr&   r'   r*   r+   ÚddÚ	__class__s         €r!   r0   zSEModule.__init__(   s•   ø€ Ø¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐÝ”9˜X x°9Ñ'<ÐRÐRÈ!ÐRÈrÐRÐRˆŒÝ”G DÐ)Ñ)Ô)ˆŒ	Ý”9˜X¨Ñ2°HÐRÐRÈ!ÐRÈrÐRÐRˆŒÝ”z‘|”|ˆŒˆˆr#   c                 óæ   — |}|                      dd¬¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }||z  S )N)é   é   T)Úkeepdim)Úmeanr1   r   r3   r5   )r6   ÚxÚmodule_inputs      r!   ÚforwardzSEModule.forward0   sc   € ØˆØ�FŠF�6 4ˆFÑ(Ô(ˆØ�HŠH�Q‰KŒKˆØ�IŠI�a‰LŒLˆØ�HŠH�Q‰KŒKˆØ�LŠL˜‰OŒOˆØ˜aÑÐr#   )NN)Ú__name__Ú
__module__Ú__qualname__Úintr0   r@   Ú__classcell__©r8   s   @r!   r%   r%   &   s_   ø€ € € € € ð$ð $ ð $°ð $ð $ð $ð $ð $ð $ð ð  ð  ð  ð  ð  ð  r#   r%   c                   ó   — e Zd ZdZd„ ZdS )Ú
BottleneckzH
    Base class for bottlenecks that implements `forward()` method.
    c                 óì  — |}|                       |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }| j        �|                      |¦  «        }|                      |¦  «        |z   }|                      |¦  «        }|S ©N)	Úconv1Úbn1r   Úconv2Úbn2Úconv3Úbn3Ú
downsampleÚ	se_module©r6   r>   ÚshortcutÚouts       r!   r@   zBottleneck.forward?   sÂ   € Øˆà�jŠj˜‰mŒmˆØ�hŠh�s‰mŒmˆØ�iŠi˜‰nŒnˆà�jŠj˜‰oŒoˆØ�hŠh�s‰mŒmˆØ�iŠi˜‰nŒnˆà�jŠj˜‰oŒoˆØ�hŠh�s‰mŒmˆàŒ?Ð&Ø—’ qÑ)Ô)ˆHà�nŠn˜SÑ!Ô! HÑ,ˆØ�iŠi˜‰nŒnˆàˆ
r#   N)rA   rB   rC   Ú__doc__r@   © r#   r!   rH   rH   :   s-   € € € € € ðð ðð ð ð ð r#   rH   c                   ó`   ‡ — e Zd ZdZdZ	 	 	 	 ddedededed	ed
eej                 fˆ fd„Z	ˆ xZ
S )ÚSEBottleneckz"
    Bottleneck for SENet154.
    é   r   NÚinplanesÚplanesÚgroupsr'   ÚstriderQ   c	           	      ó2  •— ||dœ}	t          ¦   «                              ¦   «          t          j        ||dz  fdddœ|	¤Ž| _        t          j        |dz  fi |	¤Ž| _        t          j        |dz  |dz  fd|d|ddœ|	¤Ž| _        t          j        |dz  fi |	¤Ž| _        t          j        |dz  |dz  fdddœ|	¤Ž| _	        t          j        |dz  fi |	¤Ž| _
        t          j        d	¬
¦  «        | _        t          |dz  fd|i|	¤Ž| _        || _        || _        d S )Nr)   r:   r   F©r,   r   rZ   r;   ©r,   r^   Úpaddingr]   r   Tr-   r'   ©r/   r0   r   r   rK   r   rL   rM   rN   rO   rP   r2   r   r%   rR   rQ   r^   ©r6   r[   r\   r]   r'   r^   rQ   r*   r+   r7   r8   s             €r!   r0   zSEBottleneck.__init__\   s_  ø€ ð ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐÝ”Y˜x¨°!©ÐUÀÈÐUÐUÐRTÐUÐUˆŒ
Ý”> &¨1¡*Ð3Ð3°Ð3Ð3ˆŒÝ”YØ�Q‰JØ�Q‰Jð	
ð ØØØØð	
ð 	
ð ð	
ð 	
ˆŒ
õ ”> &¨1¡*Ð3Ð3°Ð3Ð3ˆŒÝ”Y˜v¨™z¨6°A©:ÐWÀ1È5ÐWÐWÐTVÐWÐWˆŒ
Ý”> &¨1¡*Ð3Ð3°Ð3Ð3ˆŒÝ”G DÐ)Ñ)Ô)ˆŒ	Ý! &¨1¡*ÐHÐH¸	ÐHÀRÐHÐHˆŒØ$ˆŒØˆŒˆˆr#   ©r   NNN©rA   rB   rC   rV   Ú	expansionrD   r   r   ÚModuler0   rE   rF   s   @r!   rY   rY   V   s¢   ø€ € € € € ðð ð €Ið Ø.2ØØðð àðð ðð ð	ð
 ðð ðð ! ¤Ô+ðð ð ð ð ð ð ð ð ð r#   rY   c                   ó`   ‡ — e Zd ZdZdZ	 	 	 	 ddedededed	ed
eej                 fˆ fd„Z	ˆ xZ
S )ÚSEResNetBottleneckzâ
    ResNet bottleneck with a Squeeze-and-Excitation module. It follows Caffe
    implementation and uses `stride=stride` in `conv1` and not in `conv2`
    (the latter is used in the torchvision implementation of ResNet).
    rZ   r   Nr[   r\   r]   r'   r^   rQ   c	                 ó  •— ||dœ}	t          ¦   «                              ¦   «          t          j        ||fdd|dœ|	¤Ž| _        t          j        |fi |	¤Ž| _        t          j        ||fdd|ddœ|	¤Ž| _        t          j        |fi |	¤Ž| _        t          j        ||dz  fdddœ|	¤Ž| _	        t          j        |dz  fi |	¤Ž| _
        t          j        d	¬
¦  «        | _        t          |dz  fd|i|	¤Ž| _        || _        || _        d S )Nr)   r   F©r,   r   r^   r;   ©r,   rb   r]   r   rZ   r`   Tr-   r'   rc   rd   s             €r!   r0   zSEResNetBottleneck.__init__†   s3  ø€ ð ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐÝ”Y˜x¨Ð`¸QÀUÐSYÐ`Ð`Ð]_Ð`Ð`ˆŒ
Ý”> &Ð/Ð/¨BÐ/Ð/ˆŒÝ”Y˜v vÐi¸1ÀaÐPVÐ]bÐiÐiÐfhÐiÐiˆŒ
Ý”> &Ð/Ð/¨BÐ/Ð/ˆŒÝ”Y˜v v°¡zÐS¸qÀuÐSÐSÐPRÐSÐSˆŒ
Ý”> &¨1¡*Ð3Ð3°Ð3Ð3ˆŒÝ”G DÐ)Ñ)Ô)ˆŒ	Ý! &¨1¡*ÐHÐH¸	ÐHÀRÐHÐHˆŒØ$ˆŒØˆŒˆˆr#   re   rf   rF   s   @r!   rj   rj   ~   s¢   ø€ € € € € ðð ð
 €Ið Ø.2ØØðð àðð ðð ð	ð
 ðð ðð ! ¤Ô+ðð ð ð ð ð ð ð ð ð r#   rj   c                   óf   ‡ — e Zd ZdZdZ	 	 	 	 	 ddedededed	ed
eej                 defˆ fd„Z	ˆ xZ
S )ÚSEResNeXtBottleneckzI
    ResNeXt bottleneck type C with a Squeeze-and-Excitation module.
    rZ   r   Nr[   r\   r]   r'   r^   rQ   Ú
base_widthc
           	      óJ  •— ||	dœ}
t          ¦   «                              ¦   «          t          j        ||dz  z  ¦  «        |z  }t	          j        ||fddddœ|
¤Ž| _        t	          j        |fi |
¤Ž| _        t	          j        ||fd|d|ddœ|
¤Ž| _	        t	          j        |fi |
¤Ž| _
        t	          j        ||dz  fddd	œ|
¤Ž| _        t	          j        |dz  fi |
¤Ž| _        t	          j        d
¬¦  «        | _        t          |dz  fd|i|
¤Ž| _        || _        || _        d S )Nr)   é@   r   Frl   r;   ra   rZ   r`   Tr-   r'   )r/   r0   ÚmathÚfloorr   r   rK   r   rL   rM   rN   rO   rP   r2   r   r%   rR   rQ   r^   )r6   r[   r\   r]   r'   r^   rQ   rp   r*   r+   r7   Úwidthr8   s               €r!   r0   zSEResNeXtBottleneck.__init__¥   sR  ø€ ð ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐÝ”
˜6 Z°"¡_Ñ5Ñ6Ô6¸Ñ?ˆÝ”Y˜x¨ÐZ¸AÀEÐRSÐZÐZÐWYÐZÐZˆŒ
Ý”> %Ð.Ð.¨2Ð.Ð.ˆŒÝ”Y˜u eÐv¸À6ÐSTÐ]cÐjoÐvÐvÐsuÐvÐvˆŒ
Ý”> %Ð.Ð.¨2Ð.Ð.ˆŒÝ”Y˜u f¨q¡jÐR¸aÀeÐRÐRÈrÐRÐRˆŒ
Ý”> &¨1¡*Ð3Ð3°Ð3Ð3ˆŒÝ”G DÐ)Ñ)Ô)ˆŒ	Ý! &¨1¡*ÐHÐH¸	ÐHÀRÐHÐHˆŒØ$ˆŒØˆŒˆˆr#   )r   NrZ   NNrf   rF   s   @r!   ro   ro   Ÿ   s¯   ø€ € € € € ðð ð €Ið Ø.2ØØØðð àðð ðð ð	ð
 ðð ðð ! ¤Ô+ðð ðð ð ð ð ð ð ð ð ð r#   ro   c                   ób   ‡ — e Zd ZdZ	 	 	 	 ddedededededeej                 fˆ fd	„Zd
„ Z	ˆ xZ
S )ÚSEResNetBlockr   Nr[   r\   r]   r'   r^   rQ   c	                 óš  •— ||dœ}	t          ¦   «                              ¦   «          t          j        ||fdd|ddœ|	¤Ž| _        t          j        |fi |	¤Ž| _        t          j        ||fdd|ddœ|	¤Ž| _        t          j        |fi |	¤Ž| _        t          j	        d¬¦  «        | _
        t          |fd	|i|	¤Ž| _        || _        || _        d S )
Nr)   r;   r   F)r,   rb   r^   r   rm   Tr-   r'   )r/   r0   r   r   rK   r   rL   rM   rN   r2   r   r%   rR   rQ   r^   rd   s             €r!   r0   zSEResNetBlock.__init__Ã   sî   ø€ ð ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐÝ”Y˜x¨Ðk¸QÈÐRXÐ_dÐkÐkÐhjÐkÐkˆŒ
Ý”> &Ð/Ð/¨BÐ/Ð/ˆŒÝ”Y˜v vÐi¸1ÀaÐPVÐ]bÐiÐiÐfhÐiÐiˆŒ
Ý”> &Ð/Ð/¨BÐ/Ð/ˆŒÝ”G DÐ)Ñ)Ô)ˆŒ	Ý! &ÐDÐD°IÐDÀÐDÐDˆŒØ$ˆŒØˆŒˆˆr#   c                 ó˜  — |}|                       |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }| j        �|                      |¦  «        }|                      |¦  «        |z   }|                      |¦  «        }|S rJ   )rK   rL   r   rM   rN   rQ   rR   rS   s       r!   r@   zSEResNetBlock.forwardÙ   s¤   € Øˆà�jŠj˜‰mŒmˆØ�hŠh�s‰mŒmˆØ�iŠi˜‰nŒnˆà�jŠj˜‰oŒoˆØ�hŠh�s‰mŒmˆØ�iŠi˜‰nŒnˆàŒ?Ð&Ø—’ qÑ)Ô)ˆHà�nŠn˜SÑ!Ô! HÑ,ˆØ�iŠi˜‰nŒnˆàˆ
r#   re   )rA   rB   rC   rg   rD   r   r   rh   r0   r@   rE   rF   s   @r!   rw   rw   À   s¥   ø€ € € € € Ø€Ið Ø.2ØØðð àðð ðð ð	ð
 ðð ðð ! ¤Ô+ðð ð ð ð ð ð,ð ð ð ð ð ð r#   rw   c                   óV  ‡ — e Zd Z	 	 	 	 	 	 	 	 	 	 d#d
eej                 deedf         dededededede	dededede
fˆ fd„Z	 	 d$d„Zej        j        d%d„¦   «         Zej        j        d&d„¦   «         Zej        j        dej        fd„¦   «         Zd'dede
fd„Zd„ Zd%d e	fd!„Zd"„ Zˆ xZS )(r   çš™™™™™É?r;   rr   Fr   r   éè  ÚavgNÚblockÚlayers.r]   r'   Ú	drop_rateÚin_chansr[   Ú	input_3x3Údownsample_kernel_sizeÚdownsample_paddingÚnum_classesÚglobal_poolc                 óf  •— t          ¦   «                              ¦   «          ||dœ}|| _        || _        || _        || _        |rÄdt          j        |ddfddddœ|¤Žfd	t          j        d)i |¤Žfd
t          j	        d¬¦  «        fdt          j        dddddœ|¤Žfdt          j        d)i |¤Žfdt          j	        d¬¦  «        fdt          j        d|dfddddœ|¤Žfdt          j        |fi |¤Žfdt          j	        d¬¦  «        fg	}nDdt          j        ||fdddddœ|¤Žfd	t          j        |fi |¤Žfd
t          j	        d¬¦  «        fg}t          j
        t          |¦  «        ¦  «        | _        t          j        ddd¬¦  «        | _        t          |dd¬¦  «        g| _         | j        |fd|d         ||dddœ|¤Ž| _        | xj        t          d|j        z  dd¬¦  «        gz  c_         | j        |fd|d         d|||	|
dœ|¤Ž| _        | xj        t          d|j        z  dd ¬¦  «        gz  c_         | j        |fd!|d         d|||	|
dœ|¤Ž| _        | xj        t          d!|j        z  d"d#¬¦  «        gz  c_         | j        |fd$|d         d|||	|
dœ|¤Ž| _        | xj        t          d$|j        z  d%d&¬¦  «        gz  c_        d$|j        z  x| _        | _        t3          | j        | j        fd'|i|¤Ž\  | _        | _        |                      ¦   «         D ]}t;          |¦  «         Œd(S )*af  
        Parameters
        ----------
        block (nn.Module): Bottleneck class.
            - For SENet154: SEBottleneck
            - For SE-ResNet models: SEResNetBottleneck
            - For SE-ResNeXt models:  SEResNeXtBottleneck
        layers (list of ints): Number of residual blocks for 4 layers of the
            network (layer1...layer4).
        groups (int): Number of groups for the 3x3 convolution in each
            bottleneck block.
            - For SENet154: 64
            - For SE-ResNet models: 1
            - For SE-ResNeXt models:  32
        reduction (int): Reduction ratio for Squeeze-and-Excitation modules.
            - For all models: 16
        dropout_p (float or None): Drop probability for the Dropout layer.
            If `None` the Dropout layer is not used.
            - For SENet154: 0.2
            - For SE-ResNet models: None
            - For SE-ResNeXt models: None
        inplanes (int):  Number of input channels for layer1.
            - For SENet154: 128
            - For SE-ResNet models: 64
            - For SE-ResNeXt models: 64
        input_3x3 (bool): If `True`, use three 3x3 convolutions instead of
            a single 7x7 convolution in layer0.
            - For SENet154: True
            - For SE-ResNet models: False
            - For SE-ResNeXt models: False
        downsample_kernel_size (int): Kernel size for downsampling convolutions
            in layer2, layer3 and layer4.
            - For SENet154: 3
            - For SE-ResNet models: 1
            - For SE-ResNeXt models: 1
        downsample_padding (int): Padding for downsampling convolutions in
            layer2, layer3 and layer4.
            - For SENet154: 1
            - For SE-ResNet models: 0
            - For SE-ResNeXt models: 0
        num_classes (int): Number of outputs in `last_linear` layer.
            - For all models: 1000
        r)   rK   rr   r;   r:   r   F)r^   rb   r   rL   Úrelu1Tr-   rM   )rr   rr   r;   rN   Úrelu2rO   rP   Úrelu3é   ©r,   r^   rb   r   )r^   Ú	ceil_modeÚlayer0)Únum_chsr'   Úmoduler   )r\   Úblocksr]   r'   rƒ   r„   rZ   Úlayer1é€   )r\   r‘   r^   r]   r'   rƒ   r„   é   Úlayer2é   é   Úlayer3i   é    Úlayer4Ú	pool_typeN)rr   )r/   r0   r[   r…   r�   r€   r   r   r   r2   Ú
Sequentialr   rŽ   Ú	MaxPool2dÚpool0ÚdictÚfeature_infoÚ_make_layerr’   rg   r•   r˜   rš   Únum_featuresÚhead_hidden_sizer
   r†   Úlast_linearÚmodulesr"   )r6   r~   r   r]   r'   r€   r�   r[   r‚   rƒ   r„   r…   r†   r*   r+   r7   Úlayer0_modulesr    r8   s                     €r!   r0   zSENet.__init__ï   s\  ø€ õx 	‰Œ×ÒÑÔÐØ¨Ð/Ð/ˆØ ˆŒØ&ˆÔØ ˆŒØ"ˆŒØð 	à�"œ) H¨b°!Ð[¸AÀqÈuÐ[Ð[ÐXZÐ[Ð[Ð\Ø�œÐ0Ð0¨RÐ0Ð0Ð1Ø�"œ'¨$Ð/Ñ/Ô/Ð0Ø�"œ)ÐU°aÀÈÐUÐUÐRTÐUÐUÐVØ�œÐ0Ð0¨RÐ0Ð0Ð1Ø�"œ'¨$Ð/Ñ/Ô/Ð0Ø�"œ) B¨°!Ð[¸AÀqÈuÐ[Ð[ÐXZÐ[Ð[Ð\Ø�œ xÐ6Ð6°2Ð6Ð6Ð7Ø�"œ'¨$Ð/Ñ/Ô/Ð0ð
ˆNˆNð �"œ) H¨hÐmÀAÈaÐYZÐafÐmÐmÐjlÐmÐmÐnØ�œ xÐ6Ð6°2Ð6Ð6Ð7Ø�"œ'¨$Ð/Ñ/Ô/Ð0ðˆNõ
 ”m¥K°Ñ$?Ô$?Ñ@Ô@ˆŒå”\ !¨A¸Ð>Ñ>Ô>ˆŒ
Ý!¨(¸aÈÐQÑQÔQÐRˆÔØ&�dÔ&Øð	
àØ˜!”9ØØØ#$Ø ð	
ð 	
ð ð	
ð 	
ˆŒð 	ÐÔ�d¨2°´Ñ+?È1ÐU]Ð^Ñ^Ô^Ð_Ñ_ÐÔØ&�dÔ&Øð

àØ˜!”9ØØØØ#9Ø1ð

ð 

ð ð

ð 

ˆŒð 	ÐÔ�d¨3°´Ñ+@ÈAÐV^Ð_Ñ_Ô_Ð`Ñ`ÐÔØ&�dÔ&Øð

àØ˜!”9ØØØØ#9Ø1ð

ð 

ð ð

ð 

ˆŒð 	ÐÔ�d¨3°´Ñ+@ÈBÐW_Ð`Ñ`Ô`ÐaÑaÐÔØ&�dÔ&Øð

àØ˜!”9ØØØØ#9Ø1ð

ð 

ð ð

ð 

ˆŒð 	ÐÔ�d¨3°´Ñ+@ÈBÐW_Ð`Ñ`Ô`ÐaÑaÐÔØ47¸%¼/Ñ4IÐIˆÔ˜DÔ1Ý->ØÔØÔð.
ð .
ð "ð.
ð ð	.
ð .
Ñ*ˆÔ˜$Ô*ð —’‘”ð 	ð 	ˆAÝ˜‰OŒOˆOˆOð	ð 	r#   c           
      ó²  — |	|
dœ}d }|dk    s| j         ||j        z  k    rPt          j        t          j        | j         ||j        z  f|||ddœ|¤Žt          j        ||j        z  fi |¤Ž¦  «        } || j         |||||fi |¤Žg}||j        z  | _         t          d|¦  «        D ]&}|                      || j         |||fi |¤Ž¦  «         Œ't          j        |Ž S )Nr)   r   FrŒ   )r[   rg   r   rœ   r   r   ÚrangeÚappend)r6   r~   r\   r‘   r]   r'   r^   rƒ   r„   r*   r+   r7   rQ   r   Úis                  r!   r¡   zSENet._make_layer�  s,  € à¨Ð/Ð/ˆØˆ
Ø�QŠ;ˆ;˜$œ-¨6°E´OÑ+CÒCÐCÝœÝ”	Ø”M 6¨E¬OÑ#;ðQØI_Ø!Ð+=ÀEðQð QàMOðQð Qõ ”˜v¨¬Ñ7Ð>Ð>¸2Ð>Ð>ñ	ô ˆJð �%˜œ v¨v°yÀ&È*Ð[Ð[ÐXZÐ[Ð[Ð\ˆØ ¤Ñ0ˆŒÝ�q˜&Ñ!Ô!ð 	Qð 	QˆAØ�MŠM˜%˜% ¤¨v°v¸yÐOÐOÈBÐOÐOÑPÔPÐPÐPåŒ}˜fÐ%Ð%r#   c                 ó0   — t          d|rdnd¬¦  «        }|S )Nz^layer0z^layer(\d+)z^layer(\d+)\.(\d+))Ústemr‘   )rŸ   )r6   ÚcoarseÚmatchers      r!   Úgroup_matcherzSENet.group_matcher”  s#   € å˜JÀÐ/b¨~¨~ÐMbÐcÑcÔcˆØˆr#   Tc                 ó   — |r
J d¦   «         ‚d S )Nz$gradient checkpointing not supportedrW   )r6   Úenables     r!   Úset_grad_checkpointingzSENet.set_grad_checkpointing™  s   € àÐAÐAÐAÑAÔAˆzÐAÐAr#   Úreturnc                 ó   — | j         S rJ   )r¤   )r6   s    r!   Úget_classifierzSENet.get_classifier�  s   € àÔÐr#   c                 óü   — t          | ¦  «        }|| _        t          | j        | j        fd|i|¤Ž\  | _        | _        | j                             | j        ¦  «         | j                             | j        ¦  «         d S )Nr›   )r   r…   r
   r¢   r†   r¤   ÚtrainÚtraining)r6   r…   r†   r7   s       r!   Úreset_classifierzSENet.reset_classifier¡  sŽ   € Ý˜dÑ#Ô#ˆØ&ˆÔÝ->ØÔ˜tÔ/ð.Nð .NØ;Fð.NØJLð.Nð .NÑ*ˆÔ˜$Ô*àÔ×Ò˜tœ}Ñ-Ô-Ð-ØÔ×Ò˜tœ}Ñ-Ô-Ð-Ð-Ð-r#   c                 ó  — |                       |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|                      |¦  «        }|S rJ   )rŽ   rž   r’   r•   r˜   rš   ©r6   r>   s     r!   Úforward_featureszSENet.forward_features©  sa   € Ø�KŠK˜‰NŒNˆØ�JŠJ�q‰MŒMˆØ�KŠK˜‰NŒNˆØ�KŠK˜‰NŒNˆØ�KŠK˜‰NŒNˆØ�KŠK˜‰NŒNˆØˆr#   Ú
pre_logitsc                 ó¶   — |                       |¦  «        }| j        dk    r!t          j        || j        | j        ¬¦  «        }|r|n|                      |¦  «        S )Nr   )Úpr¸   )r†   r€   ÚFÚdropoutr¸   r¤   )r6   r>   r½   s      r!   Úforward_headzSENet.forward_head²  sY   € Ø×Ò˜QÑÔˆØŒ>˜BÒÐÝ”	˜!˜tœ~¸¼ÐFÑFÔFˆAØÐ7ˆqˆq D×$4Ò$4°QÑ$7Ô$7Ð7r#   c                 óZ   — |                       |¦  «        }|                      |¦  «        }|S rJ   )r¼   rÂ   r»   s     r!   r@   zSENet.forward¸  s-   € Ø×!Ò! !Ñ$Ô$ˆØ×Ò˜aÑ Ô ˆØˆr#   )
r{   r;   rr   Fr   r   r|   r}   NN)r   r   r   NN©F)T)r}   )rA   rB   rC   r   r   rh   r   rD   ÚfloatÚboolÚstrr0   r¡   ÚtorchÚjitÚignorer¯   r²   rµ   r¹   r¼   rÂ   r@   rE   rF   s   @r!   r   r   í   sü  ø€ € € € € ð  #ØØØ#Ø*+Ø&'Ø#Ø$ØØðPð Pà˜œ	”?ðPð ˜#˜s˜(”OðPð ð	Pð
 ðPð ðPð ðPð ðPð ðPð %(ðPð !$ðPð ðPð ðPð Pð Pð Pð Pð Pðd LMØW[ð&ð &ð &ð &ð& „YÔðð ð ñ Ôðð „YÔðBð Bð Bñ ÔðBð „YÔð  ¤	ð  ð  ð  ñ Ôð ð.ð .¨Cð .¸cð .ð .ð .ð .ðð ð ð8ð 8¨$ð 8ð 8ð 8ð 8ðð ð ð ð ð ð r#   Fc                 ó*   — t          t          | |fi |¤ŽS rJ   )r   r   )ÚvariantÚ
pretrainedÚkwargss      r!   Ú_create_senetrÏ   ¾  s   € Ý¥ w°
ÐEÐE¸fÐEÐEÐEr#   Ú c                 ó6   — | dddddt           t          dddd	œ|¥S )
Nr|   )r;   éà   rÒ   )r‹   r‹   g      ì?Úbilinearzlayer0.conv1r¤   z
apache-2.0)Úurlr…   Ú
input_sizeÚ	pool_sizeÚcrop_pctÚinterpolationr=   ÚstdÚ
first_convÚ
classifierÚlicenser   )rÔ   rÎ   s     r!   Ú_cfgrÝ   Â  s7   € à 4°}ÐSYØ¨JÝ%Õ.BØ$°MÈlð	ð ð
 ðð r#   zmhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/legacy_senet154-e9eb9fe6.pth)rÔ   zhhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/seresnet18-4bb0ce65.pthÚbicubic)rÔ   rØ   zhhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/seresnet34-a4004e63.pthzhhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-cadene/se_resnet50-ce0d4300.pthzihttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-cadene/se_resnet101-7e38fcc6.pthzihttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-cadene/se_resnet152-d17c99b7.pthzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/seresnext26_32x4d-65ebdb501.pthzwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/legacy_se_resnext50_32x4d-f3651bad.pthzxhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/legacy_se_resnext101_32x4d-37725eac.pth)	zlegacy_senet154.in1kzlegacy_seresnet18.in1kzlegacy_seresnet34.in1kzlegacy_seresnet50.in1kzlegacy_seresnet101.in1kzlegacy_seresnet152.in1kzlegacy_seresnext26_32x4d.in1kzlegacy_seresnext50_32x4d.in1kzlegacy_seresnext101_32x4d.in1kr³   c           	      óh   — t          t          g d¢dd¬¦  «        }t          d| fi t          |fi |¤Ž¤ŽS )N©r:   r:   r:   r:   r   r—   ©r~   r   r]   r'   Úlegacy_seresnet18©rŸ   rw   rÏ   ©rÍ   rÎ   Ú
model_argss      r!   râ   râ   ä  óO   € åÝ L L L¸ÀbðJñ Jô J€JåÐ,¨jÐWÐW½DÀÐ<VÐ<VÈvÐ<VÐ<VÐWÐWÐWr#   c           	      óh   — t          t          g d¢dd¬¦  «        }t          d| fi t          |fi |¤Ž¤ŽS )N©r;   rZ   é   r;   r   r—   rá   Úlegacy_seresnet34rã   rä   s      r!   rê   rê   ë  ræ   r#   c           	      óh   — t          t          g d¢dd¬¦  «        }t          d| fi t          |fi |¤Ž¤ŽS )Nrè   r   r—   rá   Úlegacy_seresnet50©rŸ   rj   rÏ   rä   s      r!   rì   rì   ò  sO   € åÝ ¨¨¨¸aÈ2ðOñ Oô O€JåÐ,¨jÐWÐW½DÀÐ<VÐ<VÈvÐ<VÐ<VÐWÐWÐWr#   c           	      óh   — t          t          g d¢dd¬¦  «        }t          d| fi t          |fi |¤Ž¤ŽS )N©r;   rZ   é   r;   r   r—   rá   Úlegacy_seresnet101rí   rä   s      r!   rñ   rñ   ù  óP   € åÝ ¨¨¨¸qÈBðPñ Pô P€JåÐ-¨zÐXÐX½TÀ*Ð=WÐ=WÐPVÐ=WÐ=WÐXÐXÐXr#   c           	      óh   — t          t          g d¢dd¬¦  «        }t          d| fi t          |fi |¤Ž¤ŽS )N©r;   r”   é$   r;   r   r—   rá   Úlegacy_seresnet152rí   rä   s      r!   rö   rö      rò   r#   c           
      óp   — t          t          g d¢dddddd¬¦  «        }t          d	| fi t          |fi |¤Ž¤ŽS )
Nrô   rr   r—   r;   r   r“   T)r~   r   r]   r'   rƒ   r„   r[   r‚   Úlegacy_senet154)rŸ   rY   rÏ   rä   s      r!   rø   rø     s[   € åÝ = = =¸ÀrØ °QÀ#ÐQUðWñ Wô W€Jõ Ð*¨JÐUÐU½$¸zÐ:TÐ:TÈVÐ:TÐ:TÐUÐUÐUr#   c           	      óh   — t          t          g d¢dd¬¦  «        }t          d| fi t          |fi |¤Ž¤ŽS )Nrà   r™   r—   rá   Úlegacy_seresnext26_32x4d©rŸ   ro   rÏ   rä   s      r!   rú   rú     óP   € åÝ!¨,¨,¨,¸rÈRðQñ Qô Q€JåÐ3°ZÐ^Ð^Å4È
ÐC]ÐC]ÐV\ÐC]ÐC]Ð^Ð^Ð^r#   c           	      óh   — t          t          g d¢dd¬¦  «        }t          d| fi t          |fi |¤Ž¤ŽS )Nrè   r™   r—   rá   Úlegacy_seresnext50_32x4drû   rä   s      r!   rþ   rþ     rü   r#   c           	      óh   — t          t          g d¢dd¬¦  «        }t          d| fi t          |fi |¤Ž¤ŽS )Nrï   r™   r—   rá   Úlegacy_seresnext101_32x4drû   rä   s      r!   r   r     sP   € åÝ!¨-¨-¨-ÀÈbðRñ Rô R€JåÐ4°jÐ_Ð_ÅDÈÐD^ÐD^ÐW]ÐD^ÐD^Ð_Ð_Ð_r#   rÄ   )rÐ   )/rV   rs   Úcollectionsr   Útypingr   r   r   rÈ   Útorch.nnr   Útorch.nn.functionalÚ
functionalrÀ   Ú	timm.datar   r	   Útimm.layersr
   r   Ú_builderr   Ú	_registryr   r   Ú__all__r"   rh   r%   rH   rY   rj   ro   rw   r   rÏ   rÝ   Údefault_cfgsrâ   rê   rì   rñ   rö   rø   rú   rþ   r   rW   r#   r!   ú<module>r     s7  ððð ð €€€Ø #Ð #Ð #Ð #Ð #Ð #Ø (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 à AÐ AÐ AÐ AÐ AÐ AÐ AÐ AØ ;Ð ;Ð ;Ð ;Ð ;Ð ;Ð ;Ð ;Ø *Ð *Ð *Ð *Ð *Ð *Ø <Ð <Ð <Ð <Ð <Ð <Ð <Ð <àˆ)€ð&ð &ð &ð ð  ð  ð  ð  ˆrŒyñ  ô  ð  ð(ð ð ð ð �”ñ ô ð ð8%ð %ð %ð %ð %�:ñ %ô %ð %ðPð ð ð ð ˜ñ ô ð ðBð ð ð ð ˜*ñ ô ð ðB*ð *ð *ð *ð *�B”Iñ *ô *ð *ðZNð Nð Nð Nð NˆBŒIñ Nô Nð NðbFð Fð Fð Fðð ð ð ð %Ð$Ø ˜DØ{ð}ñ }ô }à"˜dØvØð!ñ !ô !ð #˜dØvðxñ xô xà"˜dØvðxñ xô xà#˜tØwð yñ  yô  yà#˜tØwð yñ  yô  yà%) TØ~Øð&!ñ &!ô &!ð &* Tð Fð&Gñ &Gô &Gà&* dð Gð'Hñ 'Hô 'Hð'&ð &ñ ô €ð0 ðXð X°Uð Xð Xð Xñ „ðXð ðXð X°Uð Xð Xð Xñ „ðXð ðXð X°Uð Xð Xð Xñ „ðXð ðYð Y°eð Yð Yð Yñ „ðYð ðYð Y°eð Yð Yð Yñ „ðYð ðVð V°5ð Vð Vð Vñ „ðVð ð_ð _¸Eð _ð _ð _ñ „ð_ð ð_ð _¸Eð _ð _ð _ñ „ð_ð ð`ð `¸Uð `ð `ð `ñ „ð`ð `ð `r#   