o
    ôjÞ;  ã                   @   sh  d dl mZ d dlmZmZmZmZmZ d dlZd dlm	Z	m
Z
 d dlmZ ddl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 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$G dd„ dƒZ%G dd„ de	j#ƒZ&dee% de'dee de(dede&fdd„Z)d ed!d"d#œZ*G d$d%„ d%eƒZ+G d&d'„ d'eƒZ,G d(d)„ d)eƒZ-G d*d+„ d+eƒZ.eƒ ed,e+j/fd-�dd.d/œdee+ de(dede&fd0d1„ƒƒZ0eƒ ed,e,j/fd-�dd.d/œdee, de(dede&fd2d3„ƒƒZ1eƒ ed,e-j/fd-�dd.d/œdee- de(dede&fd4d5„ƒƒZ2eƒ ed,e.j/fd-�dd.d/œdee. de(dede&fd6d7„ƒƒZ3dS )8é    )Úpartial)ÚAnyÚCallableÚListÚOptionalÚSequenceN)ÚnnÚTensor)Ú
functionalé   )ÚConv2dNormActivationÚPermute)ÚStochasticDepth)ÚImageClassification)Ú_log_api_usage_onceé   )Úregister_modelÚWeightsÚWeightsEnum)Ú_IMAGENET_CATEGORIES)Ú_ovewrite_named_paramÚhandle_legacy_interface)	ÚConvNeXtÚConvNeXt_Tiny_WeightsÚConvNeXt_Small_WeightsÚConvNeXt_Base_WeightsÚConvNeXt_Large_WeightsÚconvnext_tinyÚconvnext_smallÚconvnext_baseÚconvnext_largec                   @   s   e Zd Zdedefdd„ZdS )ÚLayerNorm2dÚxÚreturnc                 C   s>   |  dddd¡}t || j| j| j| j¡}|  dddd¡}|S )Nr   r   é   r   )ÚpermuteÚFÚ
layer_normÚnormalized_shapeÚweightÚbiasÚeps©Úselfr"   © r.   úl/var/www/html/academy/tools/voices/clone/sv-venv/lib/python3.10/site-packages/torchvision/models/convnext.pyÚforward   s   zLayerNorm2d.forwardN)Ú__name__Ú
__module__Ú__qualname__r	   r0   r.   r.   r.   r/   r!      s    r!   c                
       sR   e Zd Z	ddededeedejf  ddf‡ fdd„Zd	e	de	fd
d„Z
‡  ZS )ÚCNBlockNÚlayer_scaleÚstochastic_depth_probÚ
norm_layer.r#   c                    s¬   t ƒ  ¡  |d u rttjdd�}t tj||dd|dd�tg d¢ƒ||ƒtj|d| dd	�t 	¡ tjd| |dd	�tg d
¢ƒ¡| _
t t |dd¡| ¡| _t|dƒ| _d S )Nç�íµ ÷Æ°>©r+   é   r$   T)Úkernel_sizeÚpaddingÚgroupsr*   )r   r   r$   r   é   )Úin_featuresÚout_featuresr*   )r   r$   r   r   r   Úrow)ÚsuperÚ__init__r   r   Ú	LayerNormÚ
SequentialÚConv2dr   ÚLinearÚGELUÚblockÚ	ParameterÚtorchÚonesr5   r   Ústochastic_depth)r-   Údimr5   r6   r7   ©Ú	__class__r.   r/   rC   '   s   


ù	zCNBlock.__init__Úinputc                 C   s&   | j |  |¡ }|  |¡}||7 }|S ©N)r5   rI   rM   )r-   rQ   Úresultr.   r.   r/   r0   >   s   
zCNBlock.forwardrR   )r1   r2   r3   Úfloatr   r   r   ÚModulerC   r	   r0   Ú__classcell__r.   r.   rO   r/   r4   &   s    ûýüûúr4   c                   @   s8   e Zd Zdedee deddfdd„Zdefdd	„ZdS )
ÚCNBlockConfigÚinput_channelsÚout_channelsÚ
num_layersr#   Nc                 C   s   || _ || _|| _d S rR   )rX   rY   rZ   )r-   rX   rY   rZ   r.   r.   r/   rC   G   s   
zCNBlockConfig.__init__c                 C   s>   | j jd }|d7 }|d7 }|d7 }|d7 }|jdi | j¤ŽS )Nú(zinput_channels={input_channels}z, out_channels={out_channels}z, num_layers={num_layers}ú)r.   )rP   r1   ÚformatÚ__dict__)r-   Úsr.   r.   r/   Ú__repr__Q   s   zCNBlockConfig.__repr__)r1   r2   r3   Úintr   rC   Ústrr`   r.   r.   r.   r/   rW   E   s    þýü
û
rW   c                       sŽ   e Zd Z					ddee dededed	eed
e	j
f  deed
e	j
f  deddf‡ fdd„Zdedefdd„Zdedefdd„Z‡  ZS )r   ç        r8   éè  NÚblock_settingr6   r5   Únum_classesrI   .r7   Úkwargsr#   c                    sØ  t ƒ  ¡  t| ƒ |stdƒ‚t|tƒrtdd„ |D ƒƒs!tdƒ‚|d u r't}|d u r1t	t
dd�}g }|d j}	| td|	d	d	d|d d
d�¡ tdd„ |D ƒƒ}
d}|D ]D}g }t|jƒD ]}|| |
d  }| ||j||ƒ¡ |d7 }q]| tj|Ž ¡ |jd ur˜| t ||jƒtj|j|jddd�¡¡ qTtj|Ž | _t d¡| _|d }|jd ur±|jn|j}t ||ƒt d¡t ||¡¡| _|  ¡ D ] }t|tjtjfƒrétjj|jdd� |jd urétj  |j¡ qÉd S )Nz%The block_setting should not be emptyc                 S   s   g | ]}t |tƒ‘qS r.   )Ú
isinstancerW   )Ú.0r_   r.   r.   r/   Ú
<listcomp>j   s    z%ConvNeXt.__init__.<locals>.<listcomp>z/The block_setting should be List[CNBlockConfig]r8   r9   r   r$   r>   T)r;   Ústrider<   r7   Zactivation_layerr*   c                 s   s   � | ]}|j V  qd S rR   )rZ   )ri   Úcnfr.   r.   r/   Ú	<genexpr>„   s   € z$ConvNeXt.__init__.<locals>.<genexpr>g      ð?r   r   )r;   rk   éÿÿÿÿg{®Gáz”?)Ústd)!rB   rC   r   Ú
ValueErrorrh   r   ÚallÚ	TypeErrorr4   r   r!   rX   Úappendr   ÚsumÚrangerZ   r   rE   rY   rF   ÚfeaturesÚAdaptiveAvgPool2dÚavgpoolÚFlattenrG   Ú
classifierÚmodulesÚinitÚtrunc_normal_r)   r*   Úzeros_)r-   re   r6   r5   rf   rI   r7   rg   ÚlayersZfirstconv_output_channelsZtotal_stage_blocksZstage_block_idrl   ÚstageÚ_Zsd_probZ	lastblockZlastconv_output_channelsÚmrO   r.   r/   rC   [   sp   


øÿ

þÿ€ÿÿ
€üzConvNeXt.__init__r"   c                 C   s"   |   |¡}|  |¡}|  |¡}|S rR   )rv   rx   rz   r,   r.   r.   r/   Ú_forward_impl©   s   


zConvNeXt._forward_implc                 C   s
   |   |¡S rR   )rƒ   r,   r.   r.   r/   r0   ¯   s   
zConvNeXt.forward)rc   r8   rd   NN)r1   r2   r3   r   rW   rT   ra   r   r   r   rU   r   rC   r	   rƒ   r0   rV   r.   r.   rO   r/   r   Z   s2    ùþýüûúùø	÷Nr   re   r6   ÚweightsÚprogressrg   r#   c                 K   sR   |d urt |dt|jd ƒƒ t| fd|i|¤Ž}|d ur'| |j|dd�¡ |S )Nrf   Ú
categoriesr6   T)r…   Ú
check_hash)r   ÚlenÚmetar   Úload_state_dictÚget_state_dict)re   r6   r„   r…   rg   Úmodelr.   r.   r/   Ú	_convnext³   s   r�   )é    rŽ   zNhttps://github.com/pytorch/vision/tree/main/references/classification#convnexta  
        These weights improve upon the results of the original paper by using a modified version of TorchVision's
        `new training recipe
        <https://pytorch.org/blog/how-to-train-state-of-the-art-models-using-torchvision-latest-primitives/>`_.
    )Úmin_sizer†   ÚrecipeÚ_docsc                	   @   óD   e Zd Zedeed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/convnext_tiny-983f1562.pthéà   éì   ©Ú	crop_sizeÚresize_sizeiH<´úImageNet-1Kgáz®G¡T@gÓMbX	X@©zacc@1zacc@5gmçû©ñÒ@gV-²�G[@©Ú
num_paramsÚ_metricsÚ_opsÚ
_file_size©ÚurlÚ
transformsr‰   N©	r1   r2   r3   r   r   r   Ú_COMMON_METAÚIMAGENET1K_V1ÚDEFAULTr.   r.   r.   r/   r   Ñ   ó$    ÿþÿöýr   c                	   @   r’   )r   z?https://download.pytorch.org/models/convnext_small-0c510722.pthr“   éæ   r•   iHZþr˜   g�•C‹lçT@gš™™™™)X@r™   g‘í|?5^!@gÑ"Ûù~ög@rš   rŸ   Nr¢   r.   r.   r.   r/   r   å   r¦   r   c                	   @   r’   )r   z>https://download.pytorch.org/models/convnext_base-6075fbad.pthr“   éè   r•   ihÌGr˜   g‡ÙÎ÷U@gHáz®7X@r™   gö(\�Âµ.@g/Ý$!u@rš   rŸ   Nr¢   r.   r.   r.   r/   r   ù   r¦   r   c                	   @   r’   )r   z?https://download.pytorch.org/models/convnext_large-ea097f82.pthr“   r¨   r•   i¨°Ér˜   gÑ"Ûù~U@gX9´Èv>X@r™   g‘í|?5.A@gžï§ÆK”‡@rš   rŸ   Nr¢   r.   r.   r.   r/   r     r¦   r   Ú
pretrained)r„   T)r„   r…   c                 K   óX   t  | ¡} tdddƒtdddƒtdddƒtdddƒg}| dd	¡}t||| |fi |¤ŽS )
a�  ConvNeXt Tiny model architecture from the
    `A ConvNet for the 2020s <https://arxiv.org/abs/2201.03545>`_ paper.

    Args:
        weights (:class:`~torchvision.models.convnext.ConvNeXt_Tiny_Weights`, optional): The pretrained
            weights to use. See :class:`~torchvision.models.convnext.ConvNeXt_Tiny_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.convnext.ConvNext``
            base class. Please refer to the `source code
            <https://github.com/pytorch/vision/blob/main/torchvision/models/convnext.py>`_
            for more details about this class.

    .. autoclass:: torchvision.models.ConvNeXt_Tiny_Weights
        :members:
    é`   éÀ   r$   é€  é   é	   Nr6   gš™™™™™¹?)r   ÚverifyrW   Úpopr�   ©r„   r…   rg   re   r6   r.   r.   r/   r   !  ó   




ür   c                 K   rª   )
a…  ConvNeXt Small model architecture from the
    `A ConvNet for the 2020s <https://arxiv.org/abs/2201.03545>`_ paper.

    Args:
        weights (:class:`~torchvision.models.convnext.ConvNeXt_Small_Weights`, optional): The pretrained
            weights to use. See :class:`~torchvision.models.convnext.ConvNeXt_Small_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.convnext.ConvNext``
            base class. Please refer to the `source code
            <https://github.com/pytorch/vision/blob/main/torchvision/models/convnext.py>`_
            for more details about this class.

    .. autoclass:: torchvision.models.ConvNeXt_Small_Weights
        :members:
    r«   r¬   r$   r­   r®   é   Nr6   gš™™™™™Ù?)r   r°   rW   r±   r�   r²   r.   r.   r/   r   @  ó   




ür   c                 K   rª   )
a�  ConvNeXt Base model architecture from the
    `A ConvNet for the 2020s <https://arxiv.org/abs/2201.03545>`_ paper.

    Args:
        weights (:class:`~torchvision.models.convnext.ConvNeXt_Base_Weights`, optional): The pretrained
            weights to use. See :class:`~torchvision.models.convnext.ConvNeXt_Base_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.convnext.ConvNext``
            base class. Please refer to the `source code
            <https://github.com/pytorch/vision/blob/main/torchvision/models/convnext.py>`_
            for more details about this class.

    .. autoclass:: torchvision.models.ConvNeXt_Base_Weights
        :members:
    é€   é   r$   i   i   r´   Nr6   ç      à?)r   r°   rW   r±   r�   r²   r.   r.   r/   r   a  r³   r   c                 K   rª   )
a…  ConvNeXt Large model architecture from the
    `A ConvNet for the 2020s <https://arxiv.org/abs/2201.03545>`_ paper.

    Args:
        weights (:class:`~torchvision.models.convnext.ConvNeXt_Large_Weights`, optional): The pretrained
            weights to use. See :class:`~torchvision.models.convnext.ConvNeXt_Large_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.convnext.ConvNext``
            base class. Please refer to the `source code
            <https://github.com/pytorch/vision/blob/main/torchvision/models/convnext.py>`_
            for more details about this class.

    .. autoclass:: torchvision.models.ConvNeXt_Large_Weights
        :members:
    r¬   r­   r$   r®   i   r´   Nr6   r¸   )r   r°   rW   r±   r�   r²   r.   r.   r/   r    €  rµ   r    )4Ú	functoolsr   Útypingr   r   r   r   r   rK   r   r	   Útorch.nnr
   r&   Zops.miscr   r   Zops.stochastic_depthr   Ztransforms._presetsr   Úutilsr   Ú_apir   r   r   Ú_metar   Ú_utilsr   r   Ú__all__rD   r!   rU   r4   rW   r   rT   Úboolr�   r£   r   r   r   r   r¤   r   r   r   r    r.   r.   r.   r/   Ú<module>   sŠ    Yÿþýüû
úü*ÿÿÿÿþ*ÿÿÿÿþ