o
    ôj¹A  ã                   @   s¨  d dl Z d dlmZ d dlmZ d dlmZmZmZm	Z	 d dl
Z
d dlmZ d dlm  mZ d dlm  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 d	dlmZ d	dlmZm Z  g d¢Z!G dd„ dej"ƒZ#G dd„ dej$ƒZ%G dd„ dej&ƒZ'G dd„ dej"ƒZ(dej"dede)ddfdd„Z*de+de	e+e+e+e+f de+dee de)dede(fd d!„Z,d"ed#d$d%œZ-G d&d'„ d'eƒZ.G d(d)„ d)eƒZ/G d*d+„ d+eƒZ0G d,d-„ d-eƒZ1eƒ e d.e.j2fd/�dd0d1œdee. de)dede(fd2d3„ƒƒZ3eƒ e d.e/j2fd/�dd0d1œdee/ de)dede(fd4d5„ƒƒZ4eƒ e d.e0j2fd/�dd0d1œdee0 de)dede(fd6d7„ƒƒZ5eƒ e d.e1j2fd/�dd0d1œdee1 de)dede(fd8d9„ƒƒZ6dS ):é    N)ÚOrderedDict)Úpartial)ÚAnyÚListÚOptionalÚTuple)ÚTensoré   )ÚImageClassification)Ú_log_api_usage_onceé   )Úregister_modelÚWeightsÚWeightsEnum)Ú_IMAGENET_CATEGORIES)Ú_ovewrite_named_paramÚhandle_legacy_interface)	ÚDenseNetÚDenseNet121_WeightsÚDenseNet161_WeightsÚDenseNet169_WeightsÚDenseNet201_WeightsÚdensenet121Údensenet161Údensenet169Údensenet201c                       sÎ   e Zd Z	ddedededededdf‡ fd	d
„Zdee defdd„Z	dee defdd„Z
ejjdee defdd„ƒZejjdee defdd„ƒZejjdedefdd„ƒZdedefdd„Z‡  ZS )Ú_DenseLayerFÚnum_input_featuresÚgrowth_rateÚbn_sizeÚ	drop_rateÚmemory_efficientÚreturnNc                    sŒ   t ƒ  ¡  t |¡| _tjdd�| _tj||| dddd�| _t || ¡| _	tjdd�| _
tj|| |ddddd�| _t|ƒ| _|| _d S )NT©Úinplacer   F©Úkernel_sizeÚstrideÚbiasé   ©r&   r'   Úpaddingr(   )ÚsuperÚ__init__ÚnnÚBatchNorm2dÚnorm1ÚReLUÚrelu1ÚConv2dÚconv1Únorm2Úrelu2Úconv2Úfloatr    r!   )Úselfr   r   r   r    r!   ©Ú	__class__© úl/var/www/html/academy/tools/voices/clone/sv-venv/lib/python3.10/site-packages/torchvision/models/densenet.pyr-       s   


z_DenseLayer.__init__Úinputsc                 C   s&   t  |d¡}|  |  |  |¡¡¡}|S ©Nr   )ÚtorchÚcatr4   r2   r0   )r9   r>   Zconcated_featuresÚbottleneck_outputr<   r<   r=   Úbn_function/   s   z_DenseLayer.bn_functionÚinputc                 C   s   |D ]}|j r
 dS qdS )NTF)Úrequires_grad)r9   rD   Útensorr<   r<   r=   Úany_requires_grad5   s
   ÿz_DenseLayer.any_requires_gradc                    s$   ‡ fdd„}t j|g|¢R ddiŽS )Nc                     s
   ˆ   | ¡S ©N)rC   )r>   ©r9   r<   r=   Úclosure=   s   
z7_DenseLayer.call_checkpoint_bottleneck.<locals>.closureÚuse_reentrantF)ÚcpÚ
checkpoint)r9   rD   rJ   r<   rI   r=   Úcall_checkpoint_bottleneck;   s   z&_DenseLayer.call_checkpoint_bottleneckc                 C   ó   d S rH   r<   ©r9   rD   r<   r<   r=   ÚforwardB   ó   z_DenseLayer.forwardc                 C   rO   rH   r<   rP   r<   r<   r=   rQ   F   rR   c                 C   s†   t |tƒr	|g}n|}| jr"|  |¡r"tj ¡ rtdƒ‚|  |¡}n|  	|¡}|  
|  |  |¡¡¡}| jdkrAtj|| j| jd�}|S )Nz%Memory Efficient not supported in JITr   )ÚpÚtraining)Ú
isinstancer   r!   rG   r@   ÚjitÚis_scriptingÚ	ExceptionrN   rC   r7   r6   r5   r    ÚFÚdropoutrT   )r9   rD   Zprev_featuresrB   Únew_featuresr<   r<   r=   rQ   L   s   



©F)Ú__name__Ú
__module__Ú__qualname__Úintr8   Úboolr-   r   r   rC   rG   r@   rV   ÚunusedrN   Ú_overload_methodrQ   Ú__classcell__r<   r<   r:   r=   r      s0    ÿÿÿÿÿÿþr   c                       sT   e Zd ZdZ	ddedededededed	d
f‡ fdd„Zded	efdd„Z	‡  Z
S )Ú_DenseBlockr	   FÚ
num_layersr   r   r   r    r!   r"   Nc           	         sJ   t ƒ  ¡  t|ƒD ]}t|||  ||||d�}|  d|d  |¡ q	d S )N)r   r   r    r!   zdenselayer%dr   )r,   r-   Úranger   Ú
add_module)	r9   rf   r   r   r   r    r!   ÚiÚlayerr:   r<   r=   r-   c   s   
	
ûøz_DenseBlock.__init__Úinit_featuresc                 C   s6   |g}|   ¡ D ]\}}||ƒ}| |¡ qt |d¡S r?   )ÚitemsÚappendr@   rA   )r9   rk   ÚfeaturesÚnamerj   r[   r<   r<   r=   rQ   w   s
   z_DenseBlock.forwardr\   )r]   r^   r_   Ú_versionr`   r8   ra   r-   r   rQ   rd   r<   r<   r:   r=   re   `   s&    	ùþýüûúùøre   c                       s*   e Zd Zdededdf‡ fdd„Z‡  ZS )Ú_Transitionr   Únum_output_featuresr"   Nc                    sN   t ƒ  ¡  t |¡| _tjdd�| _tj||dddd�| _tj	ddd�| _
d S )NTr#   r   Fr%   r	   )r&   r'   )r,   r-   r.   r/   Únormr1   Úrelur3   ÚconvÚ	AvgPool2dÚpool)r9   r   rr   r:   r<   r=   r-   €   s
   
z_Transition.__init__)r]   r^   r_   r`   r-   rd   r<   r<   r:   r=   rq      s    "rq   c                       sp   e Zd ZdZ							dd	ed
eeeeef dedededededdf‡ fdd„Zde	de	fdd„Z
‡  ZS )r   aK  Densenet-BC model class, based on
    `"Densely Connected Convolutional Networks" <https://arxiv.org/pdf/1608.06993.pdf>`_.

    Args:
        growth_rate (int) - how many filters to add each layer (`k` in paper)
        block_config (list of 4 ints) - how many layers in each pooling block
        num_init_features (int) - the number of filters to learn in the first convolution layer
        bn_size (int) - multiplicative factor for number of bottle neck layers
          (i.e. bn_size * k features in the bottleneck layer)
        drop_rate (float) - dropout rate after each dense layer
        num_classes (int) - number of classification classes
        memory_efficient (bool) - If True, uses checkpointing. Much more memory efficient,
          but slower. Default: *False*. See `"paper" <https://arxiv.org/pdf/1707.06990.pdf>`_.
    é    ©é   é   é   é   é@   é   r   éè  Fr   Úblock_configÚnum_init_featuresr   r    Únum_classesr!   r"   Nc                    sŠ  t ƒ  ¡  t| ƒ t tdtjd|ddddd�fdt |¡fdtjd	d
�fdtj	dddd�fgƒ¡| _
|}t|ƒD ]>\}	}
t|
|||||d�}| j
 d|	d  |¡ ||
|  }|	t|ƒd krwt||d d�}| j
 d|	d  |¡ |d }q9| j
 dt |¡¡ t ||¡| _|  ¡ D ]5}t|tjƒr�tj |j¡ q�t|tjƒr´tj |jd¡ tj |jd¡ q�t|tjƒrÂtj |jd¡ q�d S )NZconv0r)   é   r	   Fr*   Znorm0Zrelu0Tr#   Zpool0r   )r&   r'   r+   )rf   r   r   r   r    r!   zdenseblock%d)r   rr   ztransition%dZnorm5r   )r,   r-   r   r.   Ú
Sequentialr   r3   r/   r1   Ú	MaxPool2drn   Ú	enumeratere   rh   Úlenrq   ÚLinearÚ
classifierÚmodulesrU   ÚinitÚkaiming_normal_ÚweightÚ	constant_r(   )r9   r   r�   r‚   r   r    rƒ   r!   Únum_featuresri   rf   ÚblockÚtransÚmr:   r<   r=   r-   ˜   sP   
üÿÿú€€ùzDenseNet.__init__Úxc                 C   s>   |   |¡}tj|dd�}t |d¡}t |d¡}|  |¡}|S )NTr#   )r   r   r   )rn   rY   rt   Úadaptive_avg_pool2dr@   ÚflattenrŠ   )r9   r”   rn   Úoutr<   r<   r=   rQ   Ô   s   

zDenseNet.forward)rx   ry   r~   r   r   r€   F)r]   r^   r_   Ú__doc__r`   r   r8   ra   r-   r   rQ   rd   r<   r<   r:   r=   r   ˆ   s6    øþýüûúùø	÷<r   ÚmodelÚweightsÚprogressr"   c                 C   sl   t  d¡}|j|dd�}t| ¡ ƒD ]}| |¡}|r.| d¡| d¡ }|| ||< ||= q|  |¡ d S )Nz]^(.*denselayer\d+\.(?:norm|relu|conv))\.((?:[12])\.(?:weight|bias|running_mean|running_var))$T)r›   Ú
check_hashr   r	   )ÚreÚcompileÚget_state_dictÚlistÚkeysÚmatchÚgroupÚload_state_dict)r™   rš   r›   ÚpatternÚ
state_dictÚkeyÚresÚnew_keyr<   r<   r=   Ú_load_state_dictÝ   s   ÿ
€rª   r   r�   r‚   Úkwargsc                 K   sL   |d urt |dt|jd ƒƒ t| ||fi |¤Ž}|d ur$t|||d� |S )Nrƒ   Ú
categories)r™   rš   r›   )r   rˆ   Úmetar   rª   )r   r�   r‚   rš   r›   r«   r™   r<   r<   r=   Ú	_densenetð   s   r®   )é   r¯   z*https://github.com/pytorch/vision/pull/116z'These weights are ported from LuaTorch.)Úmin_sizer¬   ÚrecipeÚ_docsc                	   @   óB   e Zd Zedeedd�i e¥dddddœid	d
dœ¥d�ZeZdS )r   z<https://download.pytorch.org/models/densenet121-a639ec97.pthéà   ©Ú	crop_sizeih¿y úImageNet-1Kg²�ï§Æ›R@g‘í|?5þV@©zacc@1zacc@5gyé&1¬@g¸…ëQØ>@©Ú
num_paramsÚ_metricsÚ_opsÚ
_file_size©ÚurlÚ
transformsr­   N©	r]   r^   r_   r   r   r
   Ú_COMMON_METAÚIMAGENET1K_V1ÚDEFAULTr<   r<   r<   r=   r     ó$    
ÿþÿöýr   c                	   @   r³   )r   z<https://download.pytorch.org/models/densenet161-8d451a50.pthr´   rµ   i(£µr·   gF¶óýÔHS@g¤p=
×cW@r¸   g¶óýÔxé@gV-²�—[@r¹   r¾   NrÁ   r<   r<   r<   r=   r     rÅ   r   c                	   @   r³   )r   z<https://download.pytorch.org/models/densenet169-b2777c0a.pthr´   rµ   ihç× r·   gfffffæR@gÝ$�•3W@r¸   gáz®Gá
@g´Èv¾ŸZK@r¹   r¾   NrÁ   r<   r<   r<   r=   r   3  rÅ   r   c                	   @   r³   )r   z<https://download.pytorch.org/models/densenet201-c1103571.pthr´   rµ   ihc1r·   gÓMbX9S@gHáz®WW@r¸   gD‹lçû)@gZd;ßWS@r¹   r¾   NrÁ   r<   r<   r<   r=   r   G  rÅ   r   Ú
pretrained)rš   T)rš   r›   c                 K   ó"   t  | ¡} tddd| |fi |¤ŽS )a{  Densenet-121 model from
    `Densely Connected Convolutional Networks <https://arxiv.org/abs/1608.06993>`_.

    Args:
        weights (:class:`~torchvision.models.DenseNet121_Weights`, optional): The
            pretrained weights to use. See
            :class:`~torchvision.models.DenseNet121_Weights` below for
            more details, and possible values. By default, no pre-trained
            weights are used.
        progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True.
        **kwargs: parameters passed to the ``torchvision.models.densenet.DenseNet``
            base class. Please refer to the `source code
            <https://github.com/pytorch/vision/blob/main/torchvision/models/densenet.py>`_
            for more details about this class.

    .. autoclass:: torchvision.models.DenseNet121_Weights
        :members:
    rx   ry   r~   )r   Úverifyr®   ©rš   r›   r«   r<   r<   r=   r   [  ó   
r   c                 K   rÇ   )a{  Densenet-161 model from
    `Densely Connected Convolutional Networks <https://arxiv.org/abs/1608.06993>`_.

    Args:
        weights (:class:`~torchvision.models.DenseNet161_Weights`, optional): The
            pretrained weights to use. See
            :class:`~torchvision.models.DenseNet161_Weights` below for
            more details, and possible values. By default, no pre-trained
            weights are used.
        progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True.
        **kwargs: parameters passed to the ``torchvision.models.densenet.DenseNet``
            base class. Please refer to the `source code
            <https://github.com/pytorch/vision/blob/main/torchvision/models/densenet.py>`_
            for more details about this class.

    .. autoclass:: torchvision.models.DenseNet161_Weights
        :members:
    é0   )rz   r{   é$   r|   é`   )r   rÈ   r®   rÉ   r<   r<   r=   r   u  rÊ   r   c                 K   rÇ   )a{  Densenet-169 model from
    `Densely Connected Convolutional Networks <https://arxiv.org/abs/1608.06993>`_.

    Args:
        weights (:class:`~torchvision.models.DenseNet169_Weights`, optional): The
            pretrained weights to use. See
            :class:`~torchvision.models.DenseNet169_Weights` below for
            more details, and possible values. By default, no pre-trained
            weights are used.
        progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True.
        **kwargs: parameters passed to the ``torchvision.models.densenet.DenseNet``
            base class. Please refer to the `source code
            <https://github.com/pytorch/vision/blob/main/torchvision/models/densenet.py>`_
            for more details about this class.

    .. autoclass:: torchvision.models.DenseNet169_Weights
        :members:
    rx   )rz   r{   rx   rx   r~   )r   rÈ   r®   rÉ   r<   r<   r=   r   �  rÊ   r   c                 K   rÇ   )a{  Densenet-201 model from
    `Densely Connected Convolutional Networks <https://arxiv.org/abs/1608.06993>`_.

    Args:
        weights (:class:`~torchvision.models.DenseNet201_Weights`, optional): The
            pretrained weights to use. See
            :class:`~torchvision.models.DenseNet201_Weights` below for
            more details, and possible values. By default, no pre-trained
            weights are used.
        progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True.
        **kwargs: parameters passed to the ``torchvision.models.densenet.DenseNet``
            base class. Please refer to the `source code
            <https://github.com/pytorch/vision/blob/main/torchvision/models/densenet.py>`_
            for more details about this class.

    .. autoclass:: torchvision.models.DenseNet201_Weights
        :members:
    rx   )rz   r{   rË   rx   r~   )r   rÈ   r®   rÉ   r<   r<   r=   r   ©  rÊ   r   )7r�   Úcollectionsr   Ú	functoolsr   Útypingr   r   r   r   r@   Útorch.nnr.   Ztorch.nn.functionalÚ
functionalrY   Útorch.utils.checkpointÚutilsrM   rL   r   Ztransforms._presetsr
   r   Ú_apir   r   r   Ú_metar   Ú_utilsr   r   Ú__all__ÚModuler   Ú
ModuleDictre   r…   rq   r   ra   rª   r`   r®   rÂ   r   r   r   r   rÃ   r   r   r   r   r<   r<   r<   r=   Ú<module>   sp    A	Uÿþýüûú
ùü***.