§
    ·¯¸jØð  ã                   ó   — d 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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"m#Z#m$Z$m%Z%m&Z& ddl'm(Z( dd	l)m*Z* dd
l+m,Z,m-Z- ddl.m/Z/m0Z0m1Z1 dgZ2 G d„ dej3        ¦  «        Z4 G d„ dej3        ¦  «        Z5 G d„ dej3        ¦  «        Z6eefeefeefeefeefeefdœZ7de	ee8f         de9de:fd„Z; G d„ dej3        ¦  «        Z<�ddej3        dee8         de:ddfd„Z=d„ Z>�dd!„Z?�dd#„Z@�dd$„ZA e/i d% e@d&d'd(d¬)¦  «        “d* e@d&d'd(d¬)¦  «        “d+ e@d&d,d,¬-¦  «        “d. e@d&d,d,d/¬0¦  «        “d1 e@d2d&d(d'¬3¦  «        “d4 e@d5d&d(d'¬3¦  «        “d6 e@d7d'¬8¦  «        “d9 e@d:d&d(d'¬3¦  «        “d; e@d<d&d(d'¬3¦  «        “d= e@d>d&d(d'¬3¦  «        “d? e@d@d&d'd(d¬A¦  «        “dB e@d&d'd(d¬)¦  «        “dC e@dDd&d'd(d¬A¦  «        “dE e@dFd&d'd(d¬A¦  «        “dG e@dHd&d'd(d¬A¦  «        “dI e@d&dJdKd¬L¦  «        “dM e@d&dJdKddN¬O¦  «        “i dP e@d&dJdKddN¬O¦  «        “dQ e@d&d'dR¬S¦  «        “dT e@d&dJdKddR¬U¦  «        “dV e@d&dJdKddR¬U¦  «        “dW e@d&d'dR¬S¦  «        “dX e@d&d'dR¬S¦  «        “dY e@dZd&d(d¬3¦  «        “d[ e@d\d&d(d¬3¦  «        “d] e@d^d&d(d¬3¦  «        “d_ e@d`d&d(d¬3¦  «        “da e@dbd&d(d¬3¦  «        “dc e@ddd&d(d¬3¦  «        “de e@dfd&d(d¬3¦  «        “dg e@dhd&d(d¬3¦  «        “di e@djd&d(d¬3¦  «        “dk e@dld&dJdKddN¬m¦  «        “dn e@dod&dJdKddN¬m¦  «        “¥i dp e@dqd&dJdKddN¬m¦  «        “dr e@dsd&dJdKddN¬m¦  «        “dt e@dud&dJdKddN¬m¦  «        “dv e@dwd&dx¬y¦  «        “dz e@d{d&dx¬y¦  «        “d| e@d}d&dx¬y¦  «        “d~ e@dd&dx¬y¦  «        “d€ e@d�d&dx¬y¦  «        “d‚ eAdƒd&d(d¬3¦  «        “d„ eAd…d&dJdKddN¬m¦  «        “d† eAd‡d&d(d¬3¦  «        “dˆ eAd‰d&dJdKddN¬m¦  «        “dŠ eAd‹d&d(d¬3¦  «        “dŒ eAd�d&dJdKddN¬m¦  «        “dŽ eAd�d&d(d¬3¦  «        “d� eAd‘d&dJdKddN¬m¦  «        “d’ eAd“d&dJdKddN¬m¦  «        “¥i d” eAd•d&d–d—ddN¬m¦  «        “d˜ eAd™d&d(d'¬3¦  «        “dš eAd›d&d(d'¬3¦  «        “dœ eAd�d&d(d'¬3¦  «        “dž eAdŸd&d(d¬3¦  «        “d  eAd¡d&d(d¬3¦  «        “d¢ eAd£d&d(d¬3¦  «        “d¤ eAd¥d&d(d¬3¦  «        “d¦ eAd§d&d(d¬3¦  «        “d¨ eAd©d&d¬y¦  «        “dª eAd«d&d¬y¦  «        “d¬ eAd­d&d¬y¦  «        “d® eAd¯d&d¬y¦  «        “d° eAd±d&d¬y¦  «        “d² eAd³d&d¬y¦  «        “d´ eAdµd&d¬y¦  «        “d¶ eAd·d&d¬y¦  «        “¥i d¸ e@¦   «         “d¹ e@d&eed7dºd¬»¦  «        “d¼ e@d&eedJdKddN¬½¦  «        “d¾ e@d&eed¿dÀd¬»¦  «        “dÁ e@d&eedJdKddN¬½¦  «        “dÂ e@d&eed7dºd¬»¦  «        “dÃ e@d&eedJdKd¬»¦  «        “dÄ e@d&eed7dºd¬»¦  «        “dÅ e@d&eedJdKddN¬½¦  «        “dÆ e@d&eed7dºd¬»¦  «        “dÇ e@d&eedRd7dºd¬È¦  «        “dÉ e@d&eedRd¿dÀd¬È¦  «        “dÊ e@d&eedRdJdKddN¬Ë¦  «        “dÌ e@d&eedRdJdKddN¬Ë¦  «        “dÍ e@d&eedRd7dºd¬È¦  «        “dÎ e@d&eed7dºddÏ¬Ð¦  «        “dÑ e@d&eed7dºddÏ¬Ð¦  «        “¥i dÒ e@d&eed7dºddÏ¬Ð¦  «        “dÓ e@d&eed¿dÀddÏ¬Ð¦  «        “dÔ e@d&eed¿dÀddÏ¬Ð¦  «        “dÕ e@d&eed7dºddÖ¬Ð¦  «        “d× e@d&eed¿dÀddÖ¬Ð¦  «        “dØ e@d&eed¿dÀddÖ¬Ð¦  «        “dÙ e@d&eed7dºddÚ¬Ð¦  «        “dÛ e@d&eed7dºddÚ¬Ð¦  «        “dÜ e@d&dddÝ¬Þ¦  «        “dß e@d&dddÝ¬Þ¦  «        “dà e@d&dddÝ¬Þ¦  «        “dá e@d&dddÝ¬Þ¦  «        “dâ e@d&d7dºdddã¬ä¦  «        “då e@d&d7dºdddã¬ä¦  «        “dæ e@d&d7dºdddã¬ä¦  «        “dç e@d&d,d,dèdéd'¬»¦  «        “dê e@d&d,d,dèdéd'¬»¦  «        “¥dë e@d&d,d,dèdéd'¬»¦  «        i¥¦  «        ZBe0�dde<fdì„¦   «         ZCe0�dde<fdí„¦   «         ZDe0�dde<fdî„¦   «         ZEe0�dde<fdï„¦   «         ZFe0�dde<fdð„¦   «         ZGe0�dde<fdñ„¦   «         ZHe0�dde<fdò„¦   «         ZIe0�dde<fdó„¦   «         ZJe0�dde<fdô„¦   «         ZKe0�dde<fdõ„¦   «         ZLe0�dde<fdö„¦   «         ZMe0�dde<fd÷„¦   «         ZNe0�dde<fdø„¦   «         ZOe0�dde<fdù„¦   «         ZPe0�dde<fdú„¦   «         ZQe0�dde<fdû„¦   «         ZRe0�dde<fdü„¦   «         ZSe0�dde<fdý„¦   «         ZTe0�dde<fdþ„¦   «         ZUe0�dde<fdÿ„¦   «         ZVe0�dde<f�d „¦   «         ZWe0�dde<f�d„¦   «         ZXe0�dde<f�d„¦   «         ZYe0�dde<f�d„¦   «         ZZe0�dde<f�d„¦   «         Z[e0�dde<f�d„¦   «         Z\e0�dde<f�d„¦   «         Z]e0�dde<f�d„¦   «         Z^e0�dde<f�d„¦   «         Z_e0�dde<f�d	„¦   «         Z`e0�dde<f�d
„¦   «         Za e1ebdYd[d]d_dadkdndpdrdtdvdzd|d~d€�dœ¦  «         dS (  ax   ConvNeXt

Papers:
* `A ConvNet for the 2020s` - https://arxiv.org/pdf/2201.03545.pdf
@Article{liu2022convnet,
  author  = {Zhuang Liu and Hanzi Mao and Chao-Yuan Wu and Christoph Feichtenhofer and Trevor Darrell and Saining Xie},
  title   = {A ConvNet for the 2020s},
  journal = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year    = {2022},
}

* `ConvNeXt-V2 - Co-designing and Scaling ConvNets with Masked Autoencoders` - https://arxiv.org/abs/2301.00808
@article{Woo2023ConvNeXtV2,
  title={ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders},
  author={Sanghyun Woo, Shoubhik Debnath, Ronghang Hu, Xinlei Chen, Zhuang Liu, In So Kweon and Saining Xie},
  year={2023},
  journal={arXiv preprint arXiv:2301.00808},
}

Original code and weights from:
* https://github.com/facebookresearch/ConvNeXt, original copyright below
* https://github.com/facebookresearch/ConvNeXt-V2, original copyright below

Model defs atto, femto, pico, nano and _ols / _hnf variants are timm originals.

Modifications and additions for timm hacked together by / Copyright 2022, Ross Wightman
é    )Úpartial)ÚCallableÚDictÚListÚOptionalÚTupleÚUnionN)ÚIMAGENET_DEFAULT_MEANÚIMAGENET_DEFAULT_STDÚOPENAI_CLIP_MEANÚOPENAI_CLIP_STD)Úget_device_dtypeÚtrunc_normal_ÚAvgPool2dSameÚDropPathÚcalculate_drop_path_ratesÚMlpÚGlobalResponseNormMlpÚLayerNorm2dÚ	LayerNormÚ	RmsNorm2dÚRmsNormÚSimpleNorm2dÚ
SimpleNormÚcreate_conv2dÚget_act_layerÚget_norm_layerÚmake_divisibleÚ	to_ntupleÚNormMlpClassifierHeadÚClassifierHeadé   )Úbuild_model_with_cfg)Úfeature_take_indices)Únamed_applyÚcheckpoint_seq)Úgenerate_default_cfgsÚregister_modelÚregister_model_deprecationsÚConvNeXtc                   óf   ‡ — e Zd ZdZ	 	 	 	 ddededededdf
ˆ fd	„Zd
ej        dej        fd„Zˆ xZ	S )Ú
DownsamplezDownsample module for ConvNeXt.r"   NÚin_chsÚout_chsÚstrideÚdilationÚreturnc                 ó‚  •— ||dœ}t          ¦   «                              ¦   «          |dk    r|nd}|dk    s|dk    r4|dk    r|dk    rt          nt          j        }	 |	d|dd¬¦  «        | _        nt          j        ¦   «         | _        ||k    rt          ||dfddi|¤Ž| _        dS t          j        ¦   «         | _        dS )	zåInitialize Downsample module.

        Args:
            in_chs: Number of input channels.
            out_chs: Number of output channels.
            stride: Stride for downsampling.
            dilation: Dilation rate.
        ©ÚdeviceÚdtyper"   é   TF)Ú	ceil_modeÚcount_include_padr/   N)	ÚsuperÚ__init__r   ÚnnÚ	AvgPool2dÚpoolÚIdentityr   Úconv)Úselfr-   r.   r/   r0   r4   r5   ÚddÚ
avg_strideÚavg_pool_fnÚ	__class__s             €úd/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/timm/models/convnext.pyr:   zDownsample.__init__P   sÑ   ø€ ð" ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐØ'¨1š}˜}�V�V°!ˆ
Ø�AŠ:ˆ:˜ Aš˜Ø+5¸ª?¨?¸xÈ!º|¸|�-˜-ÕQSÔQ]ˆKØ#˜ A z¸TÐUZÐ[Ñ[Ô[ˆDŒIˆIåœ™œˆDŒIà�WÒÐÝ% f¨g°qÐIÐIÀÐIÀbÐIÐIˆDŒIˆIˆIåœ™œˆDŒIˆIˆIó    Úxc                 óZ   — |                       |¦  «        }|                      |¦  «        }|S ©úForward pass.)r=   r?   ©r@   rG   s     rE   ÚforwardzDownsample.forwardo   s%   € à�IŠI�a‰LŒLˆØ�IŠI�a‰LŒLˆØˆrF   )r"   r"   NN)
Ú__name__Ú
__module__Ú__qualname__Ú__doc__Úintr:   ÚtorchÚTensorrL   Ú__classcell__©rD   s   @rE   r,   r,   M   s¬   ø€ € € € € Ø)Ð)ð ØØØð&ð &àð&ð ð&ð ð	&ð
 ð&ð 
ð&ð &ð &ð &ð &ð &ð>˜œð ¨%¬,ð ð ð ð ð ð ð ð rF   r,   c                   óî   ‡ — e Zd ZdZ	 	 	 	 	 	 	 	 	 	 	 	 	 	 ddedee         dededeeeeef         f         dede	de	de	dee         dee
ef         dee         defˆ fd„Zdej        dej        fd„Zˆ xZS )ÚConvNeXtBlockab  ConvNeXt Block.

    There are two equivalent implementations:
      (1) DwConv -> LayerNorm (channels_first) -> 1x1 Conv -> GELU -> 1x1 Conv; all in (N, C, H, W)
      (2) DwConv -> Permute to (N, H, W, C); LayerNorm (channels_last) -> Linear -> GELU -> Linear; Permute back

    Unlike the official impl, this one allows choice of 1 or 2, 1x1 conv can be faster with appropriate
    choice of LayerNorm impl, however as model size increases the tradeoffs appear to change and nn.Linear
    is a better choice. This was observed with PyTorch 1.10 on 3090 GPU, it could change over time & w/ different HW.
    Né   r"   ©r"   r"   é   FTç�íµ ÷Æ°>Úgeluç        r-   r.   Úkernel_sizer/   r0   Ú	mlp_ratioÚconv_mlpÚ	conv_biasÚuse_grnÚls_init_valueÚ	act_layerÚ
norm_layerÚ	drop_pathc           	      ó  •— ||dœ}t          ¦   «                              ¦   «          |p|} t          d¦  «        |¦  «        }t          |¦  «        }|s|rt          nt
          }t          |	rt          nt          |¬¦  «        }|| _	        t          ||f|||d         d|dœ|¤Ž| _         ||fi |¤Ž| _         ||t          ||z  ¦  «        fd|i|¤Ž| _        |
�'t          j        |
t#          j        |fi |¤Žz  ¦  «        nd| _        ||k    s|d	k    s|d         |d	         k    rt)          ||f||d         d
œ|¤Ž| _        nt          j        ¦   «         | _        |dk    rt/          |¦  «        nt          j        ¦   «         | _        dS )a[  

        Args:
            in_chs: Block input channels.
            out_chs: Block output channels (same as in_chs if None).
            kernel_size: Depthwise convolution kernel size.
            stride: Stride of depthwise convolution.
            dilation: Tuple specifying input and output dilation of block.
            mlp_ratio: MLP expansion ratio.
            conv_mlp: Use 1x1 convolutions for MLP and a NCHW compatible norm layer if True.
            conv_bias: Apply bias for all convolution (linear) layers.
            use_grn: Use GlobalResponseNorm in MLP (from ConvNeXt-V2)
            ls_init_value: Layer-scale init values, layer-scale applied if not None.
            act_layer: Activation layer.
            norm_layer: Normalization layer (defaults to LN if not specified).
            drop_path: Stochastic depth probability.
        r3   r6   )Úuse_convr   T)r^   r/   r0   Ú	depthwiseÚbiasrd   Nr"   )r/   r0   r]   )r9   r:   r   r   r   r   r   r   r   Úuse_conv_mlpr   Úconv_dwÚnormrQ   Úmlpr;   Ú	ParameterrR   ÚonesÚgammar,   Úshortcutr>   r   rf   )r@   r-   r.   r^   r/   r0   r_   r`   ra   rb   rc   rd   re   rf   r4   r5   rA   Ú	mlp_layerrD   s                     €rE   r:   zConvNeXtBlock.__init__‚   så  ø€ ðF ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐØÐ#˜VˆØ•9˜Q‘<”< Ñ)Ô)ˆÝ! )Ñ,Ô,ˆ	Øð 	@Ø(0Ð?�˜µiˆJÝ°WÐEÕ1Ð1Å#ÐPXÐYÑYÔYˆ	Ø$ˆÔÝ$ØØð	
ð $ØØ˜a”[ØØð	
ð 	
ð ð	
ð 	
ˆŒð �J˜wÐ-Ð-¨"Ð-Ð-ˆŒ	Ø�9ØÝ�	˜GÑ#Ñ$Ô$ð
ð 
ð  ð
ð ð	
ð 
ˆŒð Q^ÐPi•R”\ -µ%´*¸WÐ2KÐ2KÈÐ2KÐ2KÑ"KÑLÔLÐLÐosˆŒ
Ø�WÒÐ ¨!¢ ¨x¸¬{¸hÀq¼kÒ/IÐ/IÝ& v¨wÐb¸vÐPXÐYZÔP[ÐbÐbÐ_aÐbÐbˆDŒMˆMåœK™MœMˆDŒMØ09¸B²°� )Ñ,Ô,Ð,ÅBÄKÁMÄMˆŒˆˆrF   rG   r1   c                 ó  — |}|                       |¦  «        }| j        r+|                      |¦  «        }|                      |¦  «        }nZ|                     dddd¦  «        }|                      |¦  «        }|                      |¦  «        }|                     dddd¦  «        }| j        �0|                     | j                             dddd¦  «        ¦  «        }|                      |¦  «        |  	                    |¦  «        z   }|S )rJ   r   r6   é   r"   Néÿÿÿÿ)
rl   rk   rm   rn   Úpermuterq   ÚmulÚreshaperf   rr   )r@   rG   rr   s      rE   rL   zConvNeXtBlock.forwardÆ   sã   € àˆØ�LŠL˜‰OŒOˆØÔð 	&Ø—	’	˜!‘”ˆAØ—’˜‘”ˆAˆAà—	’	˜!˜Q  1Ñ%Ô%ˆAØ—	’	˜!‘”ˆAØ—’˜‘”ˆAØ—	’	˜!˜Q  1Ñ%Ô%ˆAØŒ:Ð!Ø—’�d”j×(Ò(¨¨B°°1Ñ5Ô5Ñ6Ô6ˆAà�NŠN˜1ÑÔ §¢¨hÑ 7Ô 7Ñ7ˆØˆrF   )NrX   r"   rY   rZ   FTFr[   r\   Nr]   NN)rM   rN   rO   rP   rQ   r   r	   r   ÚfloatÚboolÚstrr   r:   rR   rS   rL   rT   rU   s   @rE   rW   rW   v   so  ø€ € € € € ð	ð 	ð &*Ø ØØ4:Ø Ø"Ø"Ø!Ø-1Ø.4Ø-1Ø!ØØð!BRð BRàðBRð ˜c”]ðBRð ð	BRð
 ðBRð ˜C  s¨C x¤Ð0Ô1ðBRð ðBRð ðBRð ðBRð ðBRð $ Eœ?ðBRð ˜S (˜]Ô+ðBRð ! Ô*ðBRð ðBRð BRð BRð BRð BRð BRðH˜œð ¨%¬,ð ð ð ð ð ð ð ð rF   rW   c                   óò   ‡ — e Zd ZdZ	 	 	 	 	 	 	 	 	 	 	 	 	 	 dd
edededededeeef         deee                  dede	de	de	de
eef         dee         dee         ddfˆ fd„Zdej        dej        fd„Zˆ xZS )ÚConvNeXtStagez!ConvNeXt stage (multiple blocks).rX   r6   rY   Nç      ð?FTr\   r-   r.   r^   r/   Údepthr0   Údrop_path_ratesrc   r`   ra   rb   rd   re   Únorm_layer_clr1   c                 ót  •— ||dœ}t          ¦   «                              ¦   «          d| _        ||k    s|dk    s|d         |d         k    rh|dk    s|d         |d         k    rdnd}|d         dk    rdnd}t          j         ||fi |¤Žt          ||f|||d         ||
dœ|¤Ž¦  «        | _        |}nt          j        ¦   «         | _        |pdg|z  }g }t          |¦  «        D ]?}| 	                    t          d||||d         ||         ||	|
|||	r|n|d	œ|¤Ž¦  «         |}Œ@t          j        |Ž | _        d
S )aö  Initialize ConvNeXt stage.

        Args:
            in_chs: Number of input channels.
            out_chs: Number of output channels.
            kernel_size: Kernel size for depthwise convolution.
            stride: Stride for downsampling.
            depth: Number of blocks in stage.
            dilation: Dilation rates.
            drop_path_rates: Drop path rates for each block.
            ls_init_value: Initial value for layer scale.
            conv_mlp: Use convolutional MLP.
            conv_bias: Use bias in convolutions.
            use_grn: Use global response normalization.
            act_layer: Activation layer.
            norm_layer: Normalization layer.
            norm_layer_cl: Normalization layer for channels last.
        r3   Fr"   r   r6   Úsame)r^   r/   r0   Úpaddingrj   r]   )r-   r.   r^   r0   rf   rc   r`   ra   rb   rd   re   N© )r9   r:   Úgrad_checkpointingr;   Ú
Sequentialr   Ú
downsampler>   ÚrangeÚappendrW   Úblocks)r@   r-   r.   r^   r/   r€   r0   r�   rc   r`   ra   rb   rd   re   r‚   r4   r5   rA   Úds_ksÚpadÚstage_blocksÚirD   s                         €rE   r:   zConvNeXtStage.__init__Ü   s¸  ø€ ðJ ¨Ð/Ð/ˆÝ‰Œ×ÒÑÔÐØ"'ˆÔà�WÒÐ ¨¢
 
¨h°q¬k¸XÀa¼[Ò.HÐ.HØ !š˜ x°¤{°h¸q´kÒ'AÐ'A�A�AÀqˆEØ$ QœK¨!šO˜O�&�&°ˆCÝ œmØ�
˜6Ð(Ð( RÐ(Ð(ÝØØð	ð !&Ø!Ø% aœ[ØØ"ð	ð 	ð ð	ð 	ñô ˆDŒOð ˆFˆFå œk™mœmˆDŒOà)Ð9¨b¨T°E©\ˆØˆÝ�u‘”ð 	ð 	ˆAØ×Ò¥ð !ØØØ'Ø! !œØ)¨!Ô,Ø+Ø!Ø#ØØ#Ø)1ÐD˜:˜:°}ð!ð !ð ð!ð !ñ ô ð ð ˆFˆFÝ”m \Ð2ˆŒˆˆrF   rG   c                 óÐ   — |                       |¦  «        }| j        r4t          j                             ¦   «         st          | j        |¦  «        }n|                      |¦  «        }|S rI   )r‰   r‡   rR   ÚjitÚis_scriptingr&   rŒ   rK   s     rE   rL   zConvNeXtStage.forward-  sZ   € à�OŠO˜AÑÔˆØÔ"ð 	­5¬9×+AÒ+AÑ+CÔ+Cð 	Ý˜tœ{¨AÑ.Ô.ˆAˆAà—’˜A‘”ˆAØˆrF   )rX   r6   r6   rY   Nr   FTFr\   NNNN)rM   rN   rO   rP   rQ   r   r   r   rz   r{   r	   r|   r   r:   rR   rS   rL   rT   rU   s   @rE   r~   r~   Ù   si  ø€ € € € € Ø+Ð+ð  !ØØØ(.Ø59Ø#&Ø"Ø"Ø!Ø.4Ø-1Ø04ØØð#O3ð O3àðO3ð ðO3ð ð	O3ð
 ðO3ð ðO3ð ˜C ˜H”oðO3ð & d¨5¤kÔ2ðO3ð !ðO3ð ðO3ð ðO3ð ðO3ð ˜S (˜]Ô+ðO3ð ! Ô*ðO3ð $ HÔ-ðO3ð$ 
ð%O3ð O3ð O3ð O3ð O3ð O3ðb˜œð ¨%¬,ð ð ð ð ð ð ð ð rF   r~   )Ú	layernormÚlayernorm2dÚ
simplenormÚsimplenorm2dÚrmsnormÚ	rmsnorm2dre   r`   Únorm_epsc                 óD  — | pd} | t           v r`|rt           |          d         nt           |          d         }t           |          d         } |�"t          | |¬¦  «        } t          ||¬¦  «        }n0|s
J d¦   «         ‚t          | ¦  «        } | }|�t          ||¬¦  «        }| |fS )Nr”   r   r"   )ÚepszcIf a norm_layer is specified, conv MLP must be used so all norm expect rank-4, channels-first input)Ú	_NORM_MAPr   r   )re   r`   rš   r‚   s       rE   Ú_get_norm_layersrž   A  sÊ   € ØÐ*˜{€JØ•YÐÐØ4<ÐZ�	 *Ô-¨aÔ0Ð0Å)ÈJÔBWÐXYÔBZˆÝ˜zÔ*¨1Ô-ˆ
ØÐÝ  °Ð:Ñ:Ô:ˆJÝ# M°xÐ@Ñ@Ô@ˆMøàð 	rð 	rØqñ	rô 	rˆxå# JÑ/Ô/ˆ
Ø"ˆØÐÝ# M°xÐ@Ñ@Ô@ˆMØ�}Ð$Ð$rF   c            +       ó’  ‡ — e Zd ZdZ	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 d@dededededeedf         deedf         deeeedf         f         dee	         dedede	de
dee         d e
d!e
d"e
d#eeef         d$eeeef                  d%ee	         d&e	d'e	f*ˆ fd(„Zej        j        dAd)e
d*eeeeef         f         fd+„¦   «         Zej        j        dBd,e
d*dfd-„¦   «         Zej        j        d*ej        fd.„¦   «         ZdCdedee         d*dfd/„Z	 	 	 	 	 dDd1ej        d2eeeee         f                  d3e
d4e
d5ed6e
d*eeej                 eej        eej                 f         f         fd7„Z	 	 	 dEd2eeee         f         d9e
d:e
d*ee         fd;„Zd1ej        d*ej        fd<„ZdAd1ej        d=e
d*ej        fd>„Zd1ej        d*ej        fd?„Zˆ xZS )Fr*   z|ConvNeXt model architecture.

    A PyTorch impl of : `A ConvNet for the 2020s`  - https://arxiv.org/pdf/2201.03545.pdf
    ru   éè  Úavgé    ©ru   ru   é	   ru   ©é`   éÀ   é€  é   rX   r[   ÚpatchrZ   r   FNTr\   r]   Úin_chansÚnum_classesÚglobal_poolÚoutput_strideÚdepths.ÚdimsÚkernel_sizesrc   Ú	stem_typeÚ
patch_sizeÚhead_init_scaleÚhead_norm_firstÚhead_hidden_sizer`   ra   rb   rd   re   rš   Ú	drop_rateÚdrop_path_ratec                 óH  •— t          ¦   «                              ¦   «          ||dœ}|dv sJ ‚ t          d¦  «        |¦  «        }t          |||¦  «        \  }}t	          |¦  «        }|| _        || _        || _        g | _        |	dv sJ ‚|	dk    rEt          j
        t          j        ||d         f|
|
|dœ|¤Ž ||d         fi |¤Ž¦  «        | _        |
}n˜d|	v rt          |d         d	z  ¦  «        n|d         }t          j
        t          d
t          j        ||fdd	d|dœ|¤Žd|	v r
 |¦   «         nd
t          j        ||d         fdd	d|dœ|¤Ž ||d         fi |¤Žg¦  «        Ž | _        d}t          j
        ¦   «         | _        t!          ||d¬¦  «        }g }|d         }|}d} t#          d¦  «        D ]¢}!|d	k    s|!dk    rd	nd}"||k    r|"dk    r| |"z  } d}"||"z  }| dv rdnd	}#||!         }$|                     t'          ||$f||!         |"|#| f||!         ||!         |||||||dœ|¤Ž¦  «         |$}| xj        t)          ||d|!› �¬¦  «        gz  c_        Œ£t          j
        |Ž | _        |x| _        | _        |r8|rJ ‚ || j        fi |¤Ž| _        t1          | j        |f|| j        dœ|¤Ž| _        nLt          j        ¦   «         | _        t7          | j        |f||| j        |ddœ|¤Ž| _        | j        j        | _        t9          t;          t<          |¬¦  «        | ¦  «         d
S )a  
        Args:
            in_chans: Number of input image channels.
            num_classes: Number of classes for classification head.
            global_pool: Global pooling type.
            output_stride: Output stride of network, one of (8, 16, 32).
            depths: Number of blocks at each stage.
            dims: Feature dimension at each stage.
            kernel_sizes: Depthwise convolution kernel-sizes for each stage.
            ls_init_value: Init value for Layer Scale, disabled if None.
            stem_type: Type of stem.
            patch_size: Stem patch size for patch stem.
            head_init_scale: Init scaling value for classifier weights and biases.
            head_norm_first: Apply normalization before global pool + head.
            head_hidden_size: Size of MLP hidden layer in head if not None and head_norm_first == False.
            conv_mlp: Use 1x1 conv in MLP, improves speed for small networks w/ chan last.
            conv_bias: Use bias layers w/ all convolutions.
            use_grn: Use Global Response Norm (ConvNeXt-V2) in MLP.
            act_layer: Activation layer type.
            norm_layer: Normalization layer type.
            drop_rate: Head pre-classifier dropout rate.
            drop_path_rate: Stochastic depth drop rate.
        r3   )é   é   r¢   rZ   )rª   ÚoverlapÚoverlap_tieredÚoverlap_actrª   r   )r^   r/   rj   Útieredr6   Nru   r"   )r^   r/   r…   rj   ÚactT)Ú	stagewise)r"   r6   )r^   r/   r0   r€   r�   rc   r`   ra   rb   rd   re   r‚   zstages.)Únum_chsÚ	reductionÚmodule)Ú	pool_typer·   r\   )Úhidden_sizerÅ   r·   re   rd   )r´   )r9   r:   r   rž   r   r¬   r«   r·   Úfeature_infor;   rˆ   ÚConv2dÚstemr   ÚfilterÚstagesr   rŠ   r‹   r~   ÚdictÚnum_featuresr¶   Únorm_prer!   Úheadr>   r    r%   r   Ú_init_weights)&r@   r«   r¬   r­   r®   r¯   r°   r±   rc   r²   r³   r´   rµ   r¶   r`   ra   rb   rd   re   rš   r·   r¸   r4   r5   rA   r‚   Ústem_strideÚmid_chsÚdp_ratesrË   Úprev_chsÚcurr_strider0   r�   r/   Úfirst_dilationr.   rD   s&                                        €rE   r:   zConvNeXt.__init__Y  sC  ø€ õb 	‰Œ×ÒÑÔÐØ¨Ð/Ð/ˆØ Ð+Ð+Ð+Ð+Ø#•y ‘|”| LÑ1Ô1ˆÝ$4°ZÀÈ8Ñ$TÔ$TÑ!ˆ
�MÝ! )Ñ,Ô,ˆ	à&ˆÔØ ˆŒØ"ˆŒØˆÔàÐQÐQÐQÐQÐQØ˜ÒÐåœÝ”	˜( D¨¤GÐm¸ÈJÐ]fÐmÐmÐjlÐmÐmØ�
˜4 œ7Ð)Ð) bÐ)Ð)ñô ˆDŒIð %ˆKˆKà6>À)Ð6KÐ6K•n T¨!¤W°¡\Ñ2Ô2Ð2ÐQUÐVWÔQXˆGÝœ¥v¨dÝ”	˜( GÐf¸À1ÈaÐV_ÐfÐfÐceÐfÐfØ$¨	Ð1Ð1�	�	‘”�°tÝ”	˜' 4¨¤7Ðe¸À!ÈQÐU^ÐeÐeÐbdÐeÐeØ�
˜4 œ7Ð)Ð) bÐ)Ð)ð	5ñ (ô (ð ˆDŒIð ˆKå”m‘o”oˆŒÝ,¨^¸VÈtÐTÑTÔTˆØˆØ˜”7ˆØ!ˆØˆå�q‘”ð 	gð 	gˆAØ%¨Ò*Ð*¨a°!ªe¨e�Q�Q¸ˆFØ˜mÒ+Ð+°¸²
°
Ø˜FÑ"�Ø�Ø˜6Ñ!ˆKØ"*¨fÐ"4Ð"4˜Q˜Q¸!ˆNØ˜1”gˆGØ�MŠM�-ØØðð )¨œOØØ(¨(Ð3Ø˜Q”iØ (¨¤Ø+Ø!Ø#ØØ#Ø%Ø+ðð ð ðð ñ ô ð ð" ˆHàÐÔ¥$¨xÀ;ÐWdÐabÐWdÐWdÐ"eÑ"eÔ"eÐ!fÑfÐÔÐÝ”m VÐ,ˆŒØ4<Ð<ˆÔ˜DÔ1ð ð 	;Ø'Ð'Ð'Ð'Ø&˜J tÔ'8Ð?Ð?¸BÐ?Ð?ˆDŒMÝ&ØÔ!Øðð &Øœ.ð	ð ð
 ðð ˆDŒIˆIõ œK™MœMˆDŒMÝ-ØÔ!Øð	ð -Ø%Øœ.Ø%Ø ð	ð 	ð ð	ð 	ˆDŒIð %)¤IÔ$:ˆDÔ!Ý•G�M¸?ÐKÑKÔKÈTÑRÔRÐRÐRÐRrF   Úcoarser1   c                 ó0   — t          d|rdng d¢¬¦  «        S )z¾Create regex patterns for parameter grouping.

        Args:
            coarse: Use coarse grouping.

        Returns:
            Dictionary mapping group names to regex patterns.
        z^stemz^stages\.(\d+)))z^stages\.(\d+)\.downsample)r   )z^stages\.(\d+)\.blocks\.(\d+)N)z	^norm_pre)iŸ† )rÉ   rŒ   )rÌ   )r@   r×   s     rE   Úgroup_matcherzConvNeXt.group_matcheré  s;   € õ ØØ(.ð Ð$Ð$ð 5ð 5ð 5ð
ñ 
ô 
ð 	
rF   Úenablec                 ó(   — | j         D ]	}||_        Œ
dS )zEnable or disable gradient checkpointing.

        Args:
            enable: Whether to enable gradient checkpointing.
        N)rË   r‡   )r@   rÚ   Úss      rE   Úset_grad_checkpointingzConvNeXt.set_grad_checkpointingü  s*   € ð ”ð 	*ð 	*ˆAØ#)ˆAÔ Ð ð	*ð 	*rF   c                 ó   — | j         j        S )zGet the classifier module.)rÏ   Úfc)r@   s    rE   Úget_classifierzConvNeXt.get_classifier  s   € ð ŒyŒ|ÐrF   c                 óZ   — t          | ¦  «        }|| _         | j        j        ||fi |¤Ž dS )zŸReset the classifier head.

        Args:
            num_classes: Number of classes for new classifier.
            global_pool: Global pooling type.
        N)r   r¬   rÏ   Úreset)r@   r¬   r­   rA   s       rE   Úreset_classifierzConvNeXt.reset_classifier  s=   € õ ˜dÑ#Ô#ˆØ&ˆÔØˆŒ	Œ˜ [Ð7Ð7°BÐ7Ð7Ð7Ð7Ð7rF   ÚNCHWrG   Úindicesrm   Ú
stop_earlyÚ
output_fmtÚintermediates_onlyc                 óR  — |dv s
J d¦   «         ‚g }t          t          | j        ¦  «        |¦  «        \  }}	|                      |¦  «        }t          | j        ¦  «        dz
  }
t          j                             ¦   «         s|s| j        }n| j        d|	dz   …         }t          |¦  «        D ]Z\  }} ||¦  «        }||v rF|r/||
k    r)|                     |  	                    |¦  «        ¦  «         ŒE|                     |¦  «         Œ[|r|S ||
k    r|  	                    |¦  «        }||fS )aK  Forward features that returns intermediates.

        Args:
            x: Input image tensor.
            indices: Take last n blocks if int, all if None, select matching indices if sequence.
            norm: Apply norm layer to compatible intermediates.
            stop_early: Stop iterating over blocks when last desired intermediate hit.
            output_fmt: Shape of intermediate feature outputs.
            intermediates_only: Only return intermediate features.

        Returns:
            List of intermediate features or tuple of (final features, intermediates).
        )rä   zOutput shape must be NCHW.r"   N)
r$   ÚlenrË   rÉ   rR   r’   r“   Ú	enumerater‹   rÎ   )r@   rG   rå   rm   ræ   rç   rè   ÚintermediatesÚtake_indicesÚ	max_indexÚlast_idxrË   Úfeat_idxÚstages                 rE   Úforward_intermediateszConvNeXt.forward_intermediates  sR  € ð, ˜YÐ&Ð&Ð&Ð(DÑ&Ô&Ð&ØˆÝ"6µs¸4¼;Ñ7GÔ7GÈÑ"QÔ"QÑˆ�ið �IŠI�a‰LŒLˆå�t”{Ñ#Ô# aÑ'ˆÝŒ9×!Ò!Ñ#Ô#ð 	1¨:ð 	1Ø”[ˆFˆFà”[  )¨a¡- Ô0ˆFÝ(¨Ñ0Ô0ð 	,ð 	,‰OˆH�eØ��a‘”ˆAØ˜<Ð'Ð'Øð ,˜H¨Ò0Ð0Ø!×(Ò(¨¯ª°qÑ)9Ô)9Ñ:Ô:Ð:Ð:à!×(Ò(¨Ñ+Ô+Ð+øàð 	!Ø Ð à�xÒÐØ—’˜aÑ Ô ˆAà�-ÐÐrF   r"   Ú
prune_normÚ
prune_headc                 óâ   — t          t          | j        ¦  «        |¦  «        \  }}| j        d|dz   …         | _        |rt          j        ¦   «         | _        |r|                      dd¦  «         |S )aE  Prune layers not required for specified intermediates.

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

        Returns:
            List of indices that were kept.
        Nr"   r   Ú )r$   rê   rË   r;   r>   rÎ   rã   )r@   rå   ró   rô   rí   rî   s         rE   Úprune_intermediate_layersz"ConvNeXt.prune_intermediate_layersH  sq   € õ  #7µs¸4¼;Ñ7GÔ7GÈÑ"QÔ"QÑˆ�iØ”k . 9¨q¡= .Ô1ˆŒØð 	*ÝœK™MœMˆDŒMØð 	)Ø×!Ò! ! RÑ(Ô(Ð(ØÐrF   c                 ó„   — |                       |¦  «        }|                      |¦  «        }|                      |¦  «        }|S )z/Forward pass through feature extraction layers.)rÉ   rË   rÎ   rK   s     rE   Úforward_featureszConvNeXt.forward_features`  s6   € à�IŠI�a‰LŒLˆØ�KŠK˜‰NŒNˆØ�MŠM˜!ÑÔˆØˆrF   Ú
pre_logitsc                 ó^   — |r|                       |d¬¦  «        n|                       |¦  «        S )zÊForward pass through classifier head.

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

        Returns:
            Output tensor.
        T)rú   )rÏ   )r@   rG   rú   s      rE   Úforward_headzConvNeXt.forward_headg  s.   € ð 1;ÐLˆt�yŠy˜ tˆyÑ,Ô,Ð,ÀÇ	Â	È!ÁÄÐLrF   c                 óZ   — |                       |¦  «        }|                      |¦  «        }|S rI   )rù   rü   rK   s     rE   rL   zConvNeXt.forwards  s-   € à×!Ò! !Ñ$Ô$ˆØ×Ò˜aÑ Ô ˆØˆrF   )ru   r    r¡   r¢   r£   r¥   rX   r[   rª   rZ   r   FNFTFr\   NNr]   r]   NN©F)T)N)NFFrä   F)r"   FT)rM   rN   rO   rP   rQ   r|   r   r	   r   rz   r{   r   r:   rR   r’   Úignorer   r   rÙ   rÝ   r;   ÚModulerà   rã   rS   rò   r÷   rù   rü   rL   rT   rU   s   @rE   r*   r*   S  sP  ø€ € € € € ðð ð Ø#Ø$Ø!#Ø&2Ø$7Ø89Ø-1Ø$ØØ%'Ø$)Ø.2Ø"Ø"Ø!Ø.4Ø9=Ø(,Ø!Ø$&ØØð1NSð NSàðNSð ðNSð ð	NSð
 ðNSð ˜#˜s˜(”OðNSð ˜˜S˜”/ðNSð    U¨3°¨8¤_Ð 4Ô5ðNSð $ Eœ?ðNSð ðNSð ðNSð #ðNSð "ðNSð ' sœmðNSð ðNSð  ð!NSð" ð#NSð$ ˜S (˜]Ô+ð%NSð& !  s¨H }Ô!5Ô6ð'NSð( ˜u”oð)NSð* ð+NSð, "ð-NSð NSð NSð NSð NSð NSð` „YÔð
ð 
 Dð 
°T¸#¸uÀSÈ$ÀYÔ?OÐ:OÔ5Pð 
ð 
ð 
ñ Ôð
ð$ „YÔð*ð *¨Tð *¸Tð *ð *ð *ñ Ôð*ð „YÔð ¤	ð ð ð ñ Ôðð	8ð 	8¨Cð 	8¸hÀs¼mð 	8ÐW[ð 	8ð 	8ð 	8ð 	8ð 8<ØØ$Ø$Ø',ð0 ð 0 àŒ|ð0 ð ˜e C¨¨c¬ NÔ3Ô4ð0 ð ð	0 ð
 ð0 ð ð0 ð !%ð0 ð 
ˆt�E”LÔ! 5¨¬°t¸E¼LÔ7IÐ)IÔ#JÐJÔ	Kð0 ð 0 ð 0 ð 0 ðh ./Ø$Ø#ð	ð à˜3  S¤	˜>Ô*ðð ðð ð	ð
 
ˆcŒðð ð ð ð0 %¤,ð °5´<ð ð ð ð ð
Mð 
M˜eœlð 
M¸ð 
MÈÌð 
Mð 
Mð 
Mð 
Mð˜œð ¨%¬,ð ð ð ð ð ð ð ð rF   r   rÄ   Únamer´   r1   c                 ó   — t          | t          j        ¦  «        rEt          | j        d¬¦  «         | j        �&t          j                             | j        ¦  «         dS dS t          | t          j        ¦  «        r€t          | j        d¬¦  «         t          j                             | j        ¦  «         |rFd|v rD| j        j	         
                    |¦  «         | j        j	         
                    |¦  «         dS dS dS dS )z©Initialize model weights.

    Args:
        module: Module to initialize.
        name: Module name.
        head_init_scale: Scale factor for head initialization.
    g{®Gáz”?)ÚstdNúhead.)Ú
isinstancer;   rÈ   r   Úweightrj   ÚinitÚzeros_ÚLinearÚdataÚmul_)rÄ   r  r´   s      rE   rÐ   rÐ   z  sø   € õ �&�"œ)Ñ$Ô$ð 	3Ý�f”m¨Ð-Ñ-Ô-Ð-ØŒ;Ð"ÝŒG�NŠN˜6œ;Ñ'Ô'Ð'Ð'Ð'ð #Ð"å	�F�BœIÑ	&Ô	&ð 3Ý�f”m¨Ð-Ñ-Ô-Ð-Ý
Œ�Š�v”{Ñ#Ô#Ð#Øð 	3�G˜t�O�OØŒMÔ×#Ò# OÑ4Ô4Ð4ØŒKÔ×!Ò! /Ñ2Ô2Ð2Ð2Ð2ð3ð 3ð	3ð 	3�O�OrF   c                 ó†  — d| v sd| v r| S d| v r| d         } i }d| v r¥d„ |                       ¦   «         D ¦   «         }d| v r4| d         |d<   t          j        | d         j        d         ¦  «        |d	<   nMd
| v rI| d
         |d<   | d         |d<   | d         |d<   t          j        | d         j        d         ¦  «        |d	<   |S ddl}|                       ¦   «         D �]e\  }}|                     d¦  «        rŒ|                     dd¦  «        }|                     dd|¦  «        }|                     dd|¦  «        }|                     dd¦  «        }|                     dd¦  «        }d|v rL|                     dd¦  «        }|                     dd¦  «        }|                     |j        d          ¦  «        }|                     d!d"¦  «        }|                     d#¦  «        r|                     d$d%¦  «        }|j	        d&k    r8d'|vr4| 
                    ¦   «         |         j        }|                     |¦  «        }|||<   �Œg|S )(z Remap FB checkpoints -> timm zhead.norm.weightznorm_pre.weightÚmodelzvisual.trunk.stem.0.weightc                 ól   — i | ]1\  }}|                      d ¦  «        ¯|                     d d¦  «        |“Œ2S )zvisual.trunk.rö   )Ú
startswithÚreplace)Ú.0ÚkÚvs      rE   ú
<dictcomp>z(checkpoint_filter_fn.<locals>.<dictcomp>—  sA   € ÐvÐvÐv¹$¸!¸QÐXY×XdÒXdÐetÑXuÔXuÐv�A—I’I˜o¨rÑ2Ô2°AÐvÐvÐvrF   zvisual.head.proj.weightzhead.fc.weightr   zhead.fc.biaszvisual.head.mlp.fc1.weightzhead.pre_logits.fc.weightzvisual.head.mlp.fc1.biaszhead.pre_logits.fc.biaszvisual.head.mlp.fc2.weightN)zprojectors.znorms.zdownsample_layers.0.zstem.zstages.([0-9]+).([0-9]+)zstages.\1.blocks.\2z#downsample_layers.([0-9]+).([0-9]+)zstages.\1.downsample.\2Údwconvrl   Úpwconvzmlp.fcÚgrnzgrn.betazmlp.grn.biasz	grn.gammazmlp.grn.weightrv   r  zhead.fc.znorm.rm   z	head.normr6   rÏ   )ÚitemsrR   ÚzerosÚshapeÚrer  r  Úsubry   ÚndimÚ
state_dict)r  r  Úout_dictr  r  r  Úmodel_shapes          rE   Úcheckpoint_filter_fnr!  Ž  s�  € à˜ZÐ'Ð'Ð+<À
Ð+JÐ+JØÐØ�*ÐÐØ Ô(ˆ
à€HØ# zÐ1Ð1ØvÐvÀ*×BRÒBRÑBTÔBTÐvÑvÔvˆØ$¨
Ð2Ð2Ø)3Ð4MÔ)NˆHÐ%Ñ&Ý',¤{°:Ð>WÔ3XÔ3^Ð_`Ô3aÑ'bÔ'bˆH�^Ñ$Ð$Ø)¨ZÐ7Ð7Ø4>Ð?[Ô4\ˆHÐ0Ñ1Ø2<Ð=WÔ2XˆHÐ.Ñ/Ø)3Ð4PÔ)QˆHÐ%Ñ&Ý',¤{°:Ð>ZÔ3[Ô3aÐbcÔ3dÑ'eÔ'eˆH�^Ñ$Øˆà€I€I€IØ× Ò Ñ"Ô"ð ñ ‰ˆˆ1Ø�<Š<Ð1Ñ2Ô2ð 	àØ�IŠIÐ,¨gÑ6Ô6ˆØ�FŠFÐ.Ð0FÈÑJÔJˆØ�FŠFÐ9Ð;UÐWXÑYÔYˆØ�IŠI�h 	Ñ*Ô*ˆØ�IŠI�h Ñ)Ô)ˆØ�Aˆ:ˆ:Ø—	’	˜* nÑ5Ô5ˆAØ—	’	˜+Ð'7Ñ8Ô8ˆAØ—	’	˜!œ' "œ+Ñ&Ô&ˆAØ�IŠI�g˜zÑ*Ô*ˆØ�<Š<˜Ñ Ô ð 	/Ø—	’	˜& +Ñ.Ô.ˆAØŒ6�QŠ;ˆ;˜6¨˜?˜?Ø×*Ò*Ñ,Ô,¨QÔ/Ô5ˆKØ—	’	˜+Ñ&Ô&ˆAØˆ�‰‰à€OrF   Fc                 ó¼   — |                      dd¦  «        dk    r|                     dd¦  «         t          t          | |ft          t          dd¬¦  «        d	œ|¤Ž}|S )
NÚpretrained_cfgrö   ÚfcmaeÚpretrained_strictF)r   r"   r6   ru   T)Úout_indicesÚflatten_sequential)Úpretrained_filter_fnÚfeature_cfg)ÚgetÚ
setdefaultr#   r*   r!  rÌ   )ÚvariantÚ
pretrainedÚkwargsr  s       rE   Ú_create_convnextr/  »  sy   € Ø‡z‚zÐ" BÑ'Ô'¨7Ò2Ð2ð 	×ÒÐ-¨uÑ5Ô5Ð5å Ý�'˜:ðå1Ý \ÀdÐKÑKÔKðð ð ð	ð €Eð
 €LrF   rö   c                 ó6   — | dddddt           t          dddd	œ|¥S )
Nr    ©ru   éà   r2  ©rX   rX   ç      ì?Úbicubicústem.0úhead.fcz
apache-2.0)Úurlr¬   Ú
input_sizeÚ	pool_sizeÚcrop_pctÚinterpolationÚmeanr  Ú
first_convÚ
classifierÚlicense©r
   r   ©r8  r.  s     rE   Ú_cfgrC  É  s8   € àØ¨=ÀvØ¨IÝ%Õ.BØ¨iØðð ð $*ðð rF   c                 ó<   — | dddddt           t          dddd	d
ddœ|¥S )Nr    r1  r3  r4  r5  r6  r7  zcc-by-nc-4.0zarXiv:2301.00808zGConvNeXt-V2: Co-designing and Scaling ConvNets with Masked Autoencodersz/https://github.com/facebookresearch/ConvNeXt-V2)r8  r¬   r9  r:  r;  r<  r=  r  r>  r?  r@  Ú	paper_idsÚ
paper_nameÚ
origin_urlrA  rB  s     rE   Ú_cfgv2rH  Ô  sA   € àØ¨=ÀvØ¨IÝ%Õ.BØ¨iØ!Ð0BØ_ØGð
ð 
ð ð
ð 
rF   zconvnext_tiny.in12k_ft_in1kztimm/gffffffî?)ru   é   rI  )Ú	hf_hub_idr;  Útest_input_sizeÚtest_crop_pctzconvnext_small.in12k_ft_in1kz&convnext_zepto_rms.ra4_e3600_r224_in1k)ç      à?rM  rM  )rJ  r=  r  z*convnext_zepto_rms_ols.ra4_e3600_r224_in1kgÍÌÌÌÌÌì?)rJ  r=  r  r;  zconvnext_atto.d2_in1kzrhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_atto_d2-01bb0f51.pth)r8  rJ  rK  rL  zconvnext_atto_ols.a2_in1kzvhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_atto_ols_a2-78d1c8f3.pthzconvnext_atto_rms.untrained)ru   é   rN  )rK  rL  zconvnext_femto.d1_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_femto_d1-d71d5b4c.pthzconvnext_femto_ols.d1_in1kzwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_femto_ols_d1-246bf2ed.pthzconvnext_pico.d1_in1kzrhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_pico_d1-10ad7f0d.pthzconvnext_pico_ols.d1_in1kzvhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_pico_ols_d1-611f0ca7.pth)r8  rJ  r;  rK  rL  zconvnext_nano.in12k_ft_in1kzconvnext_nano.d1h_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_nano_d1h-7eb4bdea.pthzconvnext_nano_ols.d1h_in1kzwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_nano_ols_d1h-ae424a9a.pthzconvnext_tiny_hnf.a2h_in1kzwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-rsb-weights/convnext_tiny_hnf_a2h-ab7e9df2.pthz convnext_nano.r384_in12k_ft_in1k)ru   r¨   r¨   )é   rO  )rJ  r9  r:  r;  zconvnext_tiny.in12k_ft_in1k_384Úsquash)rJ  r9  r:  r;  Ú	crop_modez convnext_small.in12k_ft_in1k_384zconvnext_nano.in12ki-.  )rJ  r;  r¬   zconvnext_nano.r384_in12k)rJ  r9  r:  r;  r¬   zconvnext_nano.r384_ad_in12kzconvnext_tiny.in12kzconvnext_small.in12kzconvnext_tiny.fb_in22k_ft_in1kzDhttps://dl.fbaipublicfiles.com/convnext/convnext_tiny_22k_1k_224.pthzconvnext_small.fb_in22k_ft_in1kzEhttps://dl.fbaipublicfiles.com/convnext/convnext_small_22k_1k_224.pthzconvnext_base.fb_in22k_ft_in1kzDhttps://dl.fbaipublicfiles.com/convnext/convnext_base_22k_1k_224.pthzconvnext_large.fb_in22k_ft_in1kzEhttps://dl.fbaipublicfiles.com/convnext/convnext_large_22k_1k_224.pthz convnext_xlarge.fb_in22k_ft_in1kzJhttps://dl.fbaipublicfiles.com/convnext/convnext_xlarge_22k_1k_224_ema.pthzconvnext_tiny.fb_in1kzDhttps://dl.fbaipublicfiles.com/convnext/convnext_tiny_1k_224_ema.pthzconvnext_small.fb_in1kzEhttps://dl.fbaipublicfiles.com/convnext/convnext_small_1k_224_ema.pthzconvnext_base.fb_in1kzDhttps://dl.fbaipublicfiles.com/convnext/convnext_base_1k_224_ema.pthzconvnext_large.fb_in1kzEhttps://dl.fbaipublicfiles.com/convnext/convnext_large_1k_224_ema.pthz"convnext_tiny.fb_in22k_ft_in1k_384zDhttps://dl.fbaipublicfiles.com/convnext/convnext_tiny_22k_1k_384.pth)r8  rJ  r9  r:  r;  rQ  z#convnext_small.fb_in22k_ft_in1k_384zEhttps://dl.fbaipublicfiles.com/convnext/convnext_small_22k_1k_384.pthz"convnext_base.fb_in22k_ft_in1k_384zDhttps://dl.fbaipublicfiles.com/convnext/convnext_base_22k_1k_384.pthz#convnext_large.fb_in22k_ft_in1k_384zEhttps://dl.fbaipublicfiles.com/convnext/convnext_large_22k_1k_384.pthz$convnext_xlarge.fb_in22k_ft_in1k_384zJhttps://dl.fbaipublicfiles.com/convnext/convnext_xlarge_22k_1k_384_ema.pthzconvnext_tiny.fb_in22kzAhttps://dl.fbaipublicfiles.com/convnext/convnext_tiny_22k_224.pthiQU  )r8  rJ  r¬   zconvnext_small.fb_in22kzBhttps://dl.fbaipublicfiles.com/convnext/convnext_small_22k_224.pthzconvnext_base.fb_in22kzAhttps://dl.fbaipublicfiles.com/convnext/convnext_base_22k_224.pthzconvnext_large.fb_in22kzBhttps://dl.fbaipublicfiles.com/convnext/convnext_large_22k_224.pthzconvnext_xlarge.fb_in22kzChttps://dl.fbaipublicfiles.com/convnext/convnext_xlarge_22k_224.pthz#convnextv2_nano.fcmae_ft_in22k_in1kzWhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_nano_22k_224_ema.ptz'convnextv2_nano.fcmae_ft_in22k_in1k_384zWhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_nano_22k_384_ema.ptz#convnextv2_tiny.fcmae_ft_in22k_in1kzWhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_tiny_22k_224_ema.ptz'convnextv2_tiny.fcmae_ft_in22k_in1k_384zWhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_tiny_22k_384_ema.ptz#convnextv2_base.fcmae_ft_in22k_in1kzWhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_base_22k_224_ema.ptz'convnextv2_base.fcmae_ft_in22k_in1k_384zWhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_base_22k_384_ema.ptz$convnextv2_large.fcmae_ft_in22k_in1kzXhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_large_22k_224_ema.ptz(convnextv2_large.fcmae_ft_in22k_in1k_384zXhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_large_22k_384_ema.ptz'convnextv2_huge.fcmae_ft_in22k_in1k_384zWhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_huge_22k_384_ema.ptz'convnextv2_huge.fcmae_ft_in22k_in1k_512zWhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im22k/convnextv2_huge_22k_512_ema.pt)ru   é   rR  )é   rS  zconvnextv2_atto.fcmae_ft_in1kzUhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im1k/convnextv2_atto_1k_224_ema.ptzconvnextv2_femto.fcmae_ft_in1kzVhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im1k/convnextv2_femto_1k_224_ema.ptzconvnextv2_pico.fcmae_ft_in1kzUhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im1k/convnextv2_pico_1k_224_ema.ptzconvnextv2_nano.fcmae_ft_in1kzUhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im1k/convnextv2_nano_1k_224_ema.ptzconvnextv2_tiny.fcmae_ft_in1kzUhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im1k/convnextv2_tiny_1k_224_ema.ptzconvnextv2_base.fcmae_ft_in1kzUhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im1k/convnextv2_base_1k_224_ema.ptzconvnextv2_large.fcmae_ft_in1kzVhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im1k/convnextv2_large_1k_224_ema.ptzconvnextv2_huge.fcmae_ft_in1kzUhttps://dl.fbaipublicfiles.com/convnext/convnextv2/im1k/convnextv2_huge_1k_224_ema.ptzconvnextv2_atto.fcmaezZhttps://dl.fbaipublicfiles.com/convnext/convnextv2/pt_only/convnextv2_atto_1k_224_fcmae.ptzconvnextv2_femto.fcmaez[https://dl.fbaipublicfiles.com/convnext/convnextv2/pt_only/convnextv2_femto_1k_224_fcmae.ptzconvnextv2_pico.fcmaezZhttps://dl.fbaipublicfiles.com/convnext/convnextv2/pt_only/convnextv2_pico_1k_224_fcmae.ptzconvnextv2_nano.fcmaezZhttps://dl.fbaipublicfiles.com/convnext/convnextv2/pt_only/convnextv2_nano_1k_224_fcmae.ptzconvnextv2_tiny.fcmaezZhttps://dl.fbaipublicfiles.com/convnext/convnextv2/pt_only/convnextv2_tiny_1k_224_fcmae.ptzconvnextv2_base.fcmaezZhttps://dl.fbaipublicfiles.com/convnext/convnextv2/pt_only/convnextv2_base_1k_224_fcmae.ptzconvnextv2_large.fcmaez[https://dl.fbaipublicfiles.com/convnext/convnextv2/pt_only/convnextv2_large_1k_224_fcmae.ptzconvnextv2_huge.fcmaezZhttps://dl.fbaipublicfiles.com/convnext/convnextv2/pt_only/convnextv2_huge_1k_224_fcmae.ptzconvnextv2_small.untrainedz/convnext_base.clip_laion2b_augreg_ft_in12k_in1k)rº   rº   )rJ  r=  r  r9  r:  r;  z3convnext_base.clip_laion2b_augreg_ft_in12k_in1k_384)rJ  r=  r  r9  r:  r;  rQ  z6convnext_large_mlp.clip_laion2b_soup_ft_in12k_in1k_320)ru   é@  rT  )é
   rU  z6convnext_large_mlp.clip_laion2b_soup_ft_in12k_in1k_384z)convnext_base.clip_laion2b_augreg_ft_in1kz,convnext_base.clip_laiona_augreg_ft_in1k_384z.convnext_large_mlp.clip_laion2b_augreg_ft_in1kz2convnext_large_mlp.clip_laion2b_augreg_ft_in1k_384z*convnext_xxlarge.clip_laion2b_soup_ft_in1kz*convnext_base.clip_laion2b_augreg_ft_in12k)rJ  r=  r  r¬   r9  r:  r;  z1convnext_large_mlp.clip_laion2b_soup_ft_in12k_320z3convnext_large_mlp.clip_laion2b_augreg_ft_in12k_384)rJ  r=  r  r¬   r9  r:  r;  rQ  z1convnext_large_mlp.clip_laion2b_soup_ft_in12k_384z+convnext_xxlarge.clip_laion2b_soup_ft_in12kzconvnext_base.clip_laion2bé€  )rJ  r=  r  r9  r:  r;  r¬   z!convnext_base.clip_laion2b_augregzconvnext_base.clip_laionazconvnext_base.clip_laiona_320z$convnext_base.clip_laiona_augreg_320z&convnext_large_mlp.clip_laion2b_augregr©   z&convnext_large_mlp.clip_laion2b_ft_320z+convnext_large_mlp.clip_laion2b_ft_soup_320z"convnext_xxlarge.clip_laion2b_soupé   z$convnext_xxlarge.clip_laion2b_rewindzconvnext_tiny.dinov3_lvd1689mzdinov3-license)rJ  r;  r¬   r@  zconvnext_small.dinov3_lvd1689mzconvnext_base.dinov3_lvd1689mzconvnext_large.dinov3_lvd1689mzconvnext_tiny.eupe_lvd1689mz#fair-noncommercial-research-license)rJ  r9  r:  r;  r¬   r@  zconvnext_small.eupe_lvd1689mzconvnext_base.eupe_lvd1689mztest_convnext.r160_in1k)ru   é    rX  )é   rY  ztest_convnext2.r160_in1kztest_convnext3.r160_in1kc           	      ó^   — t          dddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )	N©r6   r6   rZ   r6   ©r¢   é@   é€   rN  Tr–   ©r¯   r°   r`   re   Úconvnext_zepto_rmsr-  )r`  ©rÌ   r/  ©r-  r.  Ú
model_argsr  s       rE   r`  r`  Ÿ  sK   € õ ˜\Ð0BÈTÐ^jÐkÑkÔk€JÝÐgÐg¸jÐgÍDÐQ[ÐLfÐLfÐ_eÐLfÐLfÐgÐg€EØ€LrF   c           	      ó`   — t          ddddd¬¦  «        }t          d	d| it          |fi |¤Ž¤Ž}|S )
Nr[  r\  Tr–   r¾   )r¯   r°   r`   re   r²   Úconvnext_zepto_rms_olsr-  )re  ra  rb  s       rE   re  re  §  sY   € õ ØÐ"4¸tÐP\Ðhuðwñ wô w€JåÐkÐkÀ*ÐkÕPTÐU_ÐPjÐPjÐciÐPjÐPjÐkÐk€EØ€LrF   c           	      ó\   — t          ddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )N©r6   r6   é   r6   ©é(   éP   rX  rT  T©r¯   r°   r`   Úconvnext_attor-  )rm  ra  rb  s       rE   rm  rm  °  sG   € õ ˜\Ð0BÈTÐRÑRÔR€JÝÐbÐb¸ÐbÅtÈJÐGaÐGaÐZ`ÐGaÐGaÐbÐb€EØ€LrF   c           	      ó^   — t          dddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )	Nrg  ri  Tr½   ©r¯   r°   r`   r²   Úconvnext_atto_olsr-  )rp  ra  rb  s       rE   rp  rp  ¸  sK   € õ ˜\Ð0BÈTÐ]mÐnÑnÔn€JÝÐfÐf¸ZÐfÍ4ÐPZÐKeÐKeÐ^dÐKeÐKeÐfÐf€EØ€LrF   c           	      ó^   — t          dddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )	Nrg  ri  Tr™   r_  Úconvnext_atto_rmsr-  )rr  ra  rb  s       rE   rr  rr  À  sK   € õ ˜\Ð0BÈTÐ^iÐjÑjÔj€JÝÐfÐf¸ZÐfÍ4ÐPZÐKeÐKeÐ^dÐKeÐKeÐfÐf€EØ€LrF   c           	      ó\   — t          ddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nrg  ©é0   r¦   r§   r¨   Trl  Úconvnext_femtor-  )rv  ra  rb  s       rE   rv  rv  È  sG   € õ ˜\Ð0BÈTÐRÑRÔR€JÝÐcÐc¸*ÐcÍÈZÐHbÐHbÐ[aÐHbÐHbÐcÐc€EØ€LrF   c           	      ó^   — t          dddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )	Nrg  rt  Tr½   ro  Úconvnext_femto_olsr-  )rx  ra  rb  s       rE   rx  rx  Ð  sK   € õ ˜\Ð0BÈTÐ]mÐnÑnÔn€JÝÐgÐg¸jÐgÍDÐQ[ÐLfÐLfÐ_eÐLfÐLfÐgÐg€EØ€LrF   c           	      ó\   — t          ddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nrg  ©r]  r^  rN  rR  Trl  Úconvnext_picor-  )r{  ra  rb  s       rE   r{  r{  Ø  óG   € õ ˜\Ð0CÈdÐSÑSÔS€JÝÐbÐb¸ÐbÅtÈJÐGaÐGaÐZ`ÐGaÐGaÐbÐb€EØ€LrF   c           	      ó^   — t          dddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )	Nrg  rz  Tr½   ro  Úconvnext_pico_olsr-  )r~  ra  rb  s       rE   r~  r~  à  sK   € õ ˜\Ð0CÈdÐ_oÐpÑpÔp€JÝÐfÐf¸ZÐfÍ4ÐPZÐKeÐKeÐ^dÐKeÐKeÐfÐf€EØ€LrF   c           	      ó\   — t          ddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )N©r6   r6   rº   r6   ©rk  rX  rT  rV  Trl  Úconvnext_nanor-  )r‚  ra  rb  s       rE   r‚  r‚  è  r|  rF   c           	      ó^   — t          dddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )	Nr€  r�  Tr¼   ro  Úconvnext_nano_olsr-  )r„  ra  rb  s       rE   r„  r„  ð  sK   € õ ˜\Ð0CÈdÐ^gÐhÑhÔh€JÝÐfÐf¸ZÐfÍ4ÐPZÐKeÐKeÐ^dÐKeÐKeÐfÐf€EØ€LrF   c           	      ó^   — t          dddd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr£   r¥   T)r¯   r°   rµ   r`   Úconvnext_tiny_hnfr-  )r†  ra  rb  s       rE   r†  r†  ø  sL   € õ ˜\Ð0CÐUYÐdhÐiÑiÔi€JÝÐfÐf¸ZÐfÍ4ÐPZÐKeÐKeÐ^dÐKeÐKeÐfÐf€EØ€LrF   c           	      óZ   — t          dd¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr£   r¥   ©r¯   r°   Úconvnext_tinyr-  )r‰  ra  rb  s       rE   r‰  r‰     sC   € å˜\Ð0CÐDÑDÔD€JÝÐbÐb¸ÐbÅtÈJÐGaÐGaÐZ`ÐGaÐGaÐbÐb€EØ€LrF   c           	      ób   — t          g d¢g d¢¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )N©ru   ru   é   ru   r¥   rˆ  Úconvnext_smallr-  )r�  ra  rb  s       rE   r�  r�    sM   € å˜]˜]˜]Ð1DÐ1DÐ1DÐEÑEÔE€JÝÐcÐc¸*ÐcÍÈZÐHbÐHbÐ[aÐHbÐHbÐcÐc€EØ€LrF   c           	      ób   — t          g d¢g d¢¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr‹  ©r^  rN  rR  rW  rˆ  Úconvnext_baser-  )r�  ra  rb  s       rE   r�  r�    sM   € å˜]˜]˜]Ð1FÐ1FÐ1FÐGÑGÔG€JÝÐbÐb¸ÐbÅtÈJÐGaÐGaÐZ`ÐGaÐGaÐbÐb€EØ€LrF   c           	      ób   — t          g d¢g d¢¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr‹  ©r§   r¨   r©   é   rˆ  Úconvnext_larger-  )r”  ra  rb  s       rE   r”  r”    sM   € å˜]˜]˜]Ð1FÐ1FÐ1FÐGÑGÔG€JÝÐcÐc¸*ÐcÍÈZÐHbÐHbÐ[aÐHbÐHbÐcÐc€EØ€LrF   c           	      ód   — t          g d¢g d¢d¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr‹  r’  r“  )r¯   r°   r¶   Úconvnext_large_mlpr-  )r–  ra  rb  s       rE   r–  r–    sQ   € å˜]˜]˜]Ð1FÐ1FÐ1FÐY]Ð^Ñ^Ô^€JÝÐgÐg¸jÐgÍDÐQ[ÐLfÐLfÐ_eÐLfÐLfÐgÐg€EØ€LrF   c           	      ób   — t          g d¢g d¢¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr‹  )rN  rR  rW  i   rˆ  Úconvnext_xlarger-  )r˜  ra  rb  s       rE   r˜  r˜  #  sM   € å˜]˜]˜]Ð1GÐ1GÐ1GÐHÑHÔH€JÝÐdÐd¸:ÐdÍÈjÐIcÐIcÐ\bÐIcÐIcÐdÐd€EØ€LrF   c           	      óŒ   — t          g d¢g d¢|                     dd¦  «        ¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )	N)ru   rZ   é   ru   )r¨   r©   r“  i   rš   çñhãˆµøä>)r¯   r°   rš   Úconvnext_xxlarger-  )rœ  ©rÌ   Úpopr/  rb  s       rE   rœ  rœ  *  sb   € å˜]˜]˜]Ð1GÐ1GÐ1GÐRX×R\ÒR\Ð]gÐimÑRnÔRnÐoÑoÔo€JÝÐeÐe¸JÐeÍ$ÈzÐJdÐJdÐ]cÐJdÐJdÐeÐe€EØ€LrF   c           	      ó`   — t          dddd d¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nrg  ri  T©r¯   r°   rb   rc   r`   Úconvnextv2_attor-  )r¡  ra  rb  s       rE   r¡  r¡  1  sW   € õ ØÐ"4¸dÐRVÐaeðgñ gô g€JåÐdÐd¸:ÐdÍÈjÐIcÐIcÐ\bÐIcÐIcÐdÐd€EØ€LrF   c           	      ó`   — t          dddd d¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nrg  rt  Tr   Úconvnextv2_femtor-  )r£  ra  rb  s       rE   r£  r£  :  sW   € õ ØÐ"4¸dÐRVÐaeðgñ gô g€JåÐeÐe¸JÐeÍ$ÈzÐJdÐJdÐ]cÐJdÐJdÐeÐe€EØ€LrF   c           	      ó`   — t          dddd d¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nrg  rz  Tr   Úconvnextv2_picor-  )r¥  ra  rb  s       rE   r¥  r¥  C  óW   € õ ØÐ"5¸tÐSWÐbfðhñ hô h€JåÐdÐd¸:ÐdÍÈjÐIcÐIcÐ\bÐIcÐIcÐdÐd€EØ€LrF   c           	      ó`   — t          dddd d¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr€  r�  Tr   Úconvnextv2_nanor-  )r¨  ra  rb  s       rE   r¨  r¨  L  r¦  rF   c           	      ó^   — t          dddd ¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr£   r¥   T©r¯   r°   rb   rc   Úconvnextv2_tinyr-  )r«  ra  rb  s       rE   r«  r«  U  sH   € å˜\Ð0CÈTÐaeÐfÑfÔf€JÝÐdÐd¸:ÐdÍÈjÐIcÐIcÐ\bÐIcÐIcÐdÐd€EØ€LrF   c           	      óf   — t          g d¢g d¢dd ¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr‹  r¥   Trª  Úconvnextv2_smallr-  )r­  ra  rb  s       rE   r­  r­  \  sR   € å˜]˜]˜]Ð1DÐ1DÐ1DÈdÐbfÐgÑgÔg€JÝÐeÐe¸JÐeÍ$ÈzÐJdÐJdÐ]cÐJdÐJdÐeÐe€EØ€LrF   c           	      óf   — t          g d¢g d¢dd ¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr‹  r�  Trª  Úconvnextv2_baser-  )r¯  ra  rb  s       rE   r¯  r¯  c  sS   € å˜]˜]˜]Ð1FÐ1FÐ1FÐPTÐdhÐiÑiÔi€JÝÐdÐd¸:ÐdÍÈjÐIcÐIcÐ\bÐIcÐIcÐdÐd€EØ€LrF   c           	      óf   — t          g d¢g d¢dd ¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr‹  r’  Trª  Úconvnextv2_larger-  )r±  ra  rb  s       rE   r±  r±  j  sS   € å˜]˜]˜]Ð1FÐ1FÐ1FÐPTÐdhÐiÑiÔi€JÝÐeÐe¸JÐeÍ$ÈzÐJdÐJdÐ]cÐJdÐJdÐeÐe€EØ€LrF   c           	      óf   — t          g d¢g d¢dd ¬¦  «        }t          dd| it          |fi |¤Ž¤Ž}|S )Nr‹  )i`  iÀ  i€  i   Trª  Úconvnextv2_huger-  )r³  ra  rb  s       rE   r³  r³  q  sS   € å˜]˜]˜]Ð1GÐ1GÐ1GÐQUÐeiÐjÑjÔj€JÝÐdÐd¸:ÐdÍÈjÐIcÐIcÐ\bÐIcÐIcÐdÐd€EØ€LrF   c           	      óŽ   — t          g d¢g d¢|                     dd¦  «        d¬¦  «        }t          d	d| it          |fi |¤Ž¤Ž}|S )
N)r"   r6   rZ   r6   )é   r¢   ru  r]  rš   r›  Ú	gelu_tanh©r¯   r°   rš   rd   Útest_convnextr-  )r¸  r�  rb  s       rE   r¸  r¸  x  sb   € å˜\˜\˜\Ð0@Ð0@Ð0@È6Ï:Ê:ÐV`ÐbfÑKgÔKgÐs~ÐÑÔ€JÝÐbÐb¸ÐbÅtÈJÐGaÐGaÐZ`ÐGaÐGaÐbÐb€EØ€LrF   c           	      óŽ   — t          g d¢g d¢|                     dd¦  «        d¬¦  «        }t          d	d| it          |fi |¤Ž¤Ž}|S )
N©r"   r"   r"   r"   ©r¢   r]  r¦   r^  rš   r›  r¶  r·  Útest_convnext2r-  )r¼  r�  rb  s       rE   r¼  r¼    sk   € å˜\˜\˜\Ð0AÐ0AÐ0AÈFÏJÊJÐWaÐcgÑLhÔLhÐtð  Añ  Aô  A€JÝÐcÐc¸*ÐcÍÈZÐHbÐHbÐ[aÐHbÐHbÐcÐc€EØ€LrF   c           	      ó�   — t          g d¢g d¢|                     dd¦  «        dd¬¦  «        }t          d
d	| it          |fi |¤Ž¤Ž}|S )Nrº  r»  rš   r›  )rX   rY  rY  ru   Úsilu)r¯   r°   rš   r±   rd   Útest_convnext3r-  )r¿  r�  rb  s       rE   r¿  r¿  †  sr   € åØˆ|ˆ|Ð"3Ð"3Ð"3¸f¿jºjÈÐUYÑ>ZÔ>ZÐiuð  BHðIñ Iô I€JåÐcÐc¸*ÐcÍÈZÐHbÐHbÐ[aÐHbÐHbÐcÐc€EØ€LrF   )Úconvnext_tiny_in22ft1kÚconvnext_small_in22ft1kÚconvnext_base_in22ft1kÚconvnext_large_in22ft1kÚconvnext_xlarge_in22ft1kÚconvnext_tiny_384_in22ft1kÚconvnext_small_384_in22ft1kÚconvnext_base_384_in22ft1kÚconvnext_large_384_in22ft1kÚconvnext_xlarge_384_in22ft1kÚconvnext_tiny_in22kÚconvnext_small_in22kÚconvnext_base_in22kÚconvnext_large_in22kÚconvnext_xlarge_in22k)Nr   rþ   )rö   )crP   Ú	functoolsr   Útypingr   r   r   r   r   r	   rR   Útorch.nnr;   Ú	timm.datar
   r   r   r   Útimm.layersr   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r   r    r!   Ú_builderr#   Ú	_featuresr$   Ú_manipulater%   r&   Ú	_registryr'   r(   r)   Ú__all__r   r,   rW   r~   r�   r|   r{   rz   rž   r*   rÐ   r!  r/  rC  rH  Údefault_cfgsr`  re  rm  rp  rr  rv  rx  r{  r~  r‚  r„  r†  r‰  r�  r�  r”  r–  r˜  rœ  r¡  r£  r¥  r¨  r«  r­  r¯  r±  r³  r¸  r¼  r¿  rM   r†   rF   rE   ú<module>rÚ     s.  ððð ðN Ð Ð Ð Ð Ð Ø ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?à €€€Ø Ð Ð Ð Ð Ð à dÐ dÐ dÐ dÐ dÐ dÐ dÐ dÐ dÐ dÐ dÐ dðð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð ð, +Ð *Ð *Ð *Ð *Ð *Ø +Ð +Ð +Ð +Ð +Ð +Ø 4Ð 4Ð 4Ð 4Ð 4Ð 4Ð 4Ð 4Ø YÐ YÐ YÐ YÐ YÐ YÐ YÐ YÐ YÐ Yàˆ,€ð&ð &ð &ð &ð &�”ñ &ô &ð &ðR`ð `ð `ð `ð `�B”Iñ `ô `ð `ðF[ð [ð [ð [ð [�B”Iñ [ô [ð [ð~ ˜yÐ)Ø Ð+Ø Ð,Ø! :Ð.Ø˜7Ð#Ø˜WÐ%ðð €	ð%  x° }Ô!5ð %Àð %ÐQVð %ð %ð %ð %ð$dð dð dð dð dˆrŒyñ dô dð dñN	3ð 3˜"œ)ð 3¨8°C¬=ð 3ÐRWð 3Ðbfð 3ð 3ð 3ð 3ð(*ð *ð *ñZð ð ð ñð ð ð ñð ð ð ð %Ð$ð z&à! 4 4ØØ }ÀCð$Iñ $Iô $Iðz&ð
 # D DØØ }ÀCð%Iñ %Iô %Iðz&ð -¨d¨dØØ /ð/3ñ /3ô /3ðz&ð 1°$°$ØØ /Øð3ñ 3ô 3ðz&ð  ˜T˜Tð AØØ%°Tð;ñ ;ô ;ð!z&ð(    ð EØØ%°Tð";ñ ";ô ";ð)z&ð0 " 4 4à%°Tð$;ñ $;ô $;ð1z&ð6 ˜d˜dð BØØ%°Tð;ñ ;ô ;ð7z&ð> ! $ $ð FØØ%°Tð#;ñ #;ô #;ð?z&ðF ˜T˜Tð AØØ%°Tð;ñ ;ô ;ðGz&ðN    ð EØØ }ÀCð"Iñ "Iô "IðOz&ðV " 4 4ØØ }ÀCð$Iñ $Iô $IðWz&ð\ ˜d˜dð BØØ }ÀCðIñ Iô Ið]z&ðd ! $ $ð FØØ }ÀCð#Iñ #Iô #Iðez&ðl ! $ $ð FØØ }ÀCð#Iñ #Iô #Iðmz&ðt '¨¨ØØ ¨H¸sð)Dñ )Dô )Dðuz&ð| & t tØØ¨8¸sÈhð(Xñ (Xô (Xð}z&ð z&ðB '¨¨ØØ ¨H¸sÈxð)Yñ )Yô )YðCz&ðJ ˜4˜4ØØ 5ð*ñ *ô *ðKz&ðP   ØØ ¨H¸sÐPUð!Wñ !Wô !WðQz&ðV " 4 4ØØ ¨H¸sÐPUð$Wñ $Wô $WðWz&ð\ ˜4˜4ØØ 5ð*ñ *ô *ð]z&ðb ˜D˜DØØ 5ð*ñ *ô *ðcz&ðj % d dØRØØ%°Sð':ñ ':ô ':ðkz&ðr & t tØSØØ%°Sð(:ñ (:ô (:ðsz&ðz % d dØRØØ%°Sð':ñ ':ô ':ð{z&ðB & t tØSØØ%°Sð(:ñ (:ô (:ðCz&ðJ '¨¨ØXØØ%°Sð):ñ ):ô ):ðKz&ðT ˜T˜TØRØØ%°Sð:ñ :ô :ðUz&ð\ ˜d˜dØSØØ%°Sð:ñ :ô :ð]z&ðd ˜T˜TØRØØ%°Sð:ñ :ô :ðez&ðl ˜d˜dØSØØ%°Sð:ñ :ô :ðmz&ðv )¨$¨$ØRØØ ¨H¸sÈhð+Xñ +Xô +Xðwz&ð~ *¨4¨4ØSØØ ¨H¸sÈhð,Xñ ,Xô ,Xðz&ð z&ð z&ðF )¨$¨$ØRØØ ¨H¸sÈhð+Xñ +Xô +XðGz&ðN *¨4¨4ØSØØ ¨H¸sÈhð,Xñ ,Xô ,XðOz&ðV +¨D¨DØXØØ ¨H¸sÈhð-Xñ -Xô -XðWz&ð` ˜d˜dØOØØðñ ô ðaz&ðh ˜t˜tØPØØð ñ  ô  ðiz&ðp ˜d˜dØOØØðñ ô ðqz&ðx ˜t˜tØPØØð ñ  ô  ðyz&ð@   ØQØØð!ñ !ô !ðAz&ðJ *¨6¨6ØeØØ%°Sð,:ñ ,:ô ,:ðKz&ðR .¨v¨vØeØØ ¨H¸sÈhð0Xñ 0Xô 0XðSz&ðZ *¨6¨6ØeØØ%°Sð,:ñ ,:ô ,:ð[z&ðb .¨v¨vØeØØ ¨H¸sÈhð0Xñ 0Xô 0Xðcz&ðj *¨6¨6ØeØØ%°Sð,:ñ ,:ô ,:ðkz&ðr .¨v¨vØeØØ ¨H¸sÈhð0Xñ 0Xô 0Xðsz&ðz +¨F¨FØfØØ%°Sð-:ñ -:ô -:ð{z&ðB /°°ØfØØ ¨H¸sÈhð1Xñ 1Xô 1XðCz&ðJ .¨v¨vØeØØ ¨H¸sÈhð0Xñ 0Xô 0XðKz&ð z&ð z&ðR .¨v¨vØeØØ ¨H¸sÈhð0Xñ 0Xô 0XðSz&ð\ $ V VØcØØ%°Tð&;ñ &;ô &;ð]z&ðd % f fØdØØ%°Tð';ñ ';ô ';ðez&ðl $ V VØcØØ%°Tð&;ñ &;ô &;ðmz&ðt $ V VØcØØ%°Sð&:ñ &:ô &:ðuz&ð| $ V VØcØØ%°Sð&:ñ &:ô &:ð}z&ðD $ V VØcØØ%°Sð&:ñ &:ô &:ðEz&ðL % f fØdØØ%°Sð':ñ ':ô ':ðMz&ðT $ V VØcØØ%°Sð&:ñ &:ô &:ðUz&ð^ ˜V˜VØhØØðñ ô ð_z&ðf ˜f˜fØiØØðñ ô ðgz&ðn ˜V˜VØhØØðñ ô ðoz&ðv ˜V˜VØhØØðñ ô ðwz&ð~ ˜V˜VØhØØðñ ô ðz&ðF ˜V˜VØhØØðñ ô ðGz&ðN ˜f˜fØiØØðñ ô ðOz&ðV ˜V˜VØhØØðñ ô ðWz&ð z&ð z&ð` ! $ $¡&¤&ðaz&ðf 6°t°tØØ ?Ø ¨F¸Sð8Bñ 8Bô 8Bðgz&ðn :¸4¸4ØØ ?Ø ¨H¸sÈhð<Xñ <Xô <Xðoz&ðv =¸d¸dØØ ?Ø ¨H¸sð?Dñ ?Dô ?Dðwz&ð~ =¸d¸dØØ ?Ø ¨H¸sÈhð?Xñ ?Xô ?Xðz&ðH	 0°°ØØ ?Ø ¨F¸Sð2Bñ 2Bô 2BðI	z&ðP	 3°D°DØØ ?Ø ¨H¸sð5Dñ 5Dô 5DðQ	z&ðX	 5°d°dØØ ?Ø ¨F¸Sð7ñ 7ô 7ðY	z&ðb	 9¸$¸$ØØ ?Ø ¨H¸sÈhð;ñ ;ô ;ðc	z&ðl	 1°$°$ØØ ?Ø ¨F¸Sð3Bñ 3Bô 3Bðm	z&ðv	 1°$°$ØØ ?ÀØ ¨F¸Sð3Bñ 3Bô 3Bðw	z&ð~	 8¸¸ØØ ?ÀØ ¨H¸sð:Dñ :Dô :Dð	z&ðF
 :¸4¸4ØØ ?ÀØ ¨H¸sÈhð<Xñ <Xô <XðG
z&ðN
 8¸¸ØØ ?ÀØ ¨H¸sÈhð:Xñ :Xô :XðO
z&ðV
 2°4°4ØØ ?ÀØ ¨F¸Sð4Bñ 4Bô 4BðW
z&ðb
 ! $ $ØØ ?Ø ¨F¸SÈcð#Sñ #Sô #Sðc
z&ðj
 (¨¨ØØ ?Ø ¨F¸SÈcð*Sñ *Sô *Sðk
z&ð z&ð z&ðr
    ØØ ?Ø ¨F¸SÈcð"Sñ "Sô "Sðs
z&ðz
 $ T TØØ ?Ø ¨H¸sÐPSð&Uñ &Uô &Uð{
z&ðB +¨D¨DØØ ?Ø ¨H¸sÐPSð-Uñ -Uô -UðCz&ðJ -¨d¨dØØ ?Ø ¨F¸SÈcð/Sñ /Sô /SðKz&ðR -¨d¨dØØ ?Ø ¨H¸sÐPSð/Uñ /Uô /UðSz&ðZ 2°4°4ØØ ?Ø ¨H¸sÐPSð4Uñ 4Uô 4Uð[z&ðb )¨$¨$ØØ ?Ø ¨F¸SÈdð+Tñ +Tô +Tðcz&ðj +¨D¨DØØ ?Ø ¨F¸SÈdð-Tñ -Tô -Tðkz&ðx $ T TØØØØ ð	&ñ &ô &ðyz&ðD % d dØØØØ ð	'ñ 'ô 'ðEz&ðP $ T TØØØØ ð	&ñ &ô &ðQz&ð\ % d dØØØØ ð	'ñ 'ô 'ð]z&ðh " 4 4ØØ ØØØØ5ð$ñ $ô $ðiz&ðx # D DØØ ØØØØ5ð%ñ %ô %ðyz&ðH " 4 4ØØ ØØØØ5ð$ñ $ô $ðIz&ðZ ˜t˜tØØ /Ø ¨F¸Tð Cñ  Cô  Cð[z&ðb   ØØ /Ø ¨F¸Tð!Cñ !Cô !Cðcz&ð z&ðj   ØØ /Ø ¨F¸Tð!Cñ !Cô !Cðkz&ð z&ñ zô z€ðz ñð °hð ð ð ñ „ðð ñð ¸(ð ð ð ñ „ðð ñð °ð ð ð ñ „ðð ñð °Xð ð ð ñ „ðð ñð °Xð ð ð ñ „ðð ñð °(ð ð ð ñ „ðð ñð °hð ð ð ñ „ðð ñð °ð ð ð ñ „ðð ñð °Xð ð ð ñ „ðð ñð °ð ð ð ñ „ðð ñð °Xð ð ð ñ „ðð ñð °Xð ð ð ñ „ðð ñð °ð ð ð ñ „ðð ñð °(ð ð ð ñ „ðð ñð °ð ð ð ñ „ðð ñð °(ð ð ð ñ „ðð ñð °hð ð ð ñ „ðð ñð °8ð ð ð ñ „ðð ñð °Hð ð ð ñ „ðð ñð °8ð ð ð ñ „ðð ñð °Hð ñ ð ñ „ðð ñð °8ð ñ ð ñ „ðð ñð °8ð ñ ð ñ „ðð ñð °8ð ñ ð ñ „ðð ñð °Hð ñ ð ñ „ðð ñð °8ð ñ ð ñ „ðð ñð °Hð ñ ð ñ „ðð ñð °8ð ñ ð ñ „ðð ñð °ð ñ ð ñ „ðð ñð °(ð ñ ð ñ „ðð ñð °(ð ñ ð ñ „ðð Ð ˜HØ>Ø@Ø>Ø@Ø BØ"FØ#HØ"FØ#HØ$JØ3Ø5Ø3Ø5Ø7ñ'ð 'ñ ô ð ð ð rF   