§
    ·¯¸j\í 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c mZ ddlmZmZmZmZ ddlmZmZmZmZmZmZmZmZ ddlmZm Z  dd	l!m"Z" dd
l#m$Z$m%Z%m&Z&m'Z'm(Z(m)Z)m*Z*m+Z+ ddl,m-Z-m.Z.m/Z/ ddl0m1Z1m2Z2 ddl3m4Z4m5Z5m6Z6 ddgZ7 G d„ dej8        ¦  «        Z9 G d„ dej8        ¦  «        Z:�dd„Z;�dd„Z<�dd„Z=�dd„Z>	 	 �dd„Z?	 	 �dd„Z@�dd„ZA�dd„ZB	 	 �dd„ZC	 �dd„ZD	 �dd„ZE�d d „ZF	 �dd!„ZG	 �dd"„ZH	 �dd#„ZI	 �dd$„ZJ	 �dd%„ZK	 	 �d!d&„ZL�dd'„ZM�d d(„ZN�d d)„ZO�d d*„ZP�d d+„ZQ�d"d-„ZR e4i d. eR¦   «         “d/ eR¦   «         “d0 eRd1d2¬3¦  «        “d4 eR¦   «         “d5 eR¦   «         “d6 eRd7d2¬3¦  «        “d8 eRd9d2¬3¦  «        “d: eR¦   «         “d; eRd<d2¬3¦  «        “d= eRd2eed>d?¬@¦  «        “dA eRd2eed>d?¬@¦  «        “dB eRd2eedCd>d¬D¦  «        “dE eR¦   «         “dF eRdGd2dH¬I¦  «        “dJ eR¦   «         “dK eRdLd2¬3¦  «        “dM eRdNd2¬3¦  «        “i dO eRdPd2¬3¦  «        “dQ eRdRd2¬3¦  «        “dS eRdTd2dU¬I¦  «        “dV eRdWd2dU¬I¦  «        “dX eRdYd2¬3¦  «        “dZ eRd2eedCd>d¬D¦  «        “d[ eRd2eed\dCd]d^d¬_¦  «        “d` eRdad2d>d¬b¦  «        “dc eRddd2d>d]d^d¬e¦  «        “df eRdgd2d^dhdid¬e¦  «        “dj eRdkd2didldmd¬e¦  «        “dn eRd2dodpddq¬r¦  «        “ds eRd2dtdud?dv¬w¦  «        “dx eRd,dydzd{¬|¦  «        “d} eRd,d~dd€¬|¦  «        “d� eRd,d‚dƒd„¬|¦  «        “d… eRd,d†d‡dˆ¬|¦  «        “¥i d‰ eR¦   «         “dŠ eR¦   «         “d‹ eR¦   «         “dŒ eRd^dhdid¬�¦  «        “dŽ eRd^dhdid¬�¦  «        “d� eR¦   «         “d� eRd2dodpddqd‘¬’¦  «        “d“ eRd,d^dhd?¬|¦  «        “d” eRd2dodpddqd‘¬’¦  «        “d• eRd–d2¬3¦  «        “d— eRd˜d2d\d]d™¬š¦  «        “d› eRdœd2d�dldž¬š¦  «        “dŸ eRd d2¬3¦  «        “d¡ eRd¢d2d�dldž¬š¦  «        “d£ eR¦   «         “d¤ eR¦   «         “d¥ eRd\d]d™¬¦¦  «        “¥i d§ eRd¨d2¬3¦  «        “d© eRd\d]d™¬¦¦  «        “dª eRd«dhd¬¬¦¦  «        “d­ eRd�dldž¬¦¦  «        “d® eRd¯d°d±¬¦¦  «        “d² eRd³d2d\d]d™ee¬´¦  «        “dµ eRd¶d2d«dhd¬ee¬´¦  «        “d· eRd¸d2d�dldžee¬´¦  «        “d¹ eRdºd2d»d^d¼d¬½¦  «        “d¾ eRd¿d2d»d^d¼d¬½¦  «        “dÀ eRdÁd2d^dmdhd¬½¦  «        “dÂ eRdÃd2didtdld¬½¦  «        “dÄ eRd^dmdhd¬Å¦  «        “dÆ eRdidtdld¬Å¦  «        “dÇ eRdmdÈd°d¬Å¦  «        “dÉ eRdmdÊd°d¬Å¦  «        “dË eRdÌd2d»¬Í¦  «        “¥i dÎ eRdÏd2d\d]d™¬š¦  «        “dÐ eRdÑd2d«dhd¬¬š¦  «        “dÒ eRdÓd2d�dldž¬š¦  «        “dÔ eRdÕd2d¯d°d±¬š¦  «        “dÖ eRd×d2dØdÙdÚ¬š¦  «        “dÛ eRdÜd2dydzd{¬š¦  «        “dÝ eRdÞd2d~dd€¬š¦  «        “dß eRdàd2dádÙdâ¬š¦  «        “dã eRdäd2d†d‡då¬š¦  «        “dæ eRdçd2eed»¬è¦  «        “dé eRdêd2eed\d]d™¬ë¦  «        “dì eRdíd2eed«dhd¬¬ë¦  «        “dî eRdïd2eed�dldž¬ë¦  «        “dð eRdñd2eed¯d°d±¬ë¦  «        “dò eRdód2eedØdÙdÚ¬ë¦  «        “dô eRdõd2eedydzd{¬ë¦  «        “dö eRd÷d2eed~dd€¬ë¦  «        “¥i dø eRdùd2eed‚dƒd„¬ë¦  «        “dú eRdûd2dØdÙdÚ¬š¦  «        “dü eRdýd2d~dd€¬š¦  «        “dþ eRdÿd2d‚dƒd„¬š¦  «        “�d  eR�dd2d»¬Í¦  «        “�d eR�dd2d\d]d™¬š¦  «        “�d eR�dd2d«dhd¬¬š¦  «        “�d eR�dd2d�dldž¬š¦  «        “�d eR�d	d2d¯d°d±¬š¦  «        “�d
 eR�dd2dØdÙdÚ¬š¦  «        “�d eR�dd2dydzd{¬š¦  «        “�d eR�dd2d~dd€¬š¦  «        “�d eR�dd2d»¬Í¦  «        “�d eR�dd2d\d]d™¬š¦  «        “�d eR�dd2d«dhd¬¬š¦  «        “�d eR�dd2d�dldž¬š¦  «        “�d eR�dd2d¯d°d±¬š¦  «        “¥i �d eR�dd2dØdÙdÚ¬š¦  «        “�d eR�dd2�d�dd»¬è¦  «        “�d eR�d d2�d�dd\d]d™¬ë¦  «        “�d! eR�d"d2�d�dd�dldž¬ë¦  «        “�d# eR�d$d2ee�¬%¦  «        “�d& eR�d'd2ee�¬%¦  «        “�d( eR�d)d2eed\d]d™¬ë¦  «        “�d* eR�d+d2�d�ddH�¬,¦  «        “�d- eR�d.d2�d�dd\d]d™dH�¬/¦  «        “�d0 eR�d1d2�d�dd«dhd¬dH�¬/¦  «        “�d2 eR�d3d2�d�dd�dldždU�¬/¦  «        “�d4 eR�d5d2�d�dd¯d°�d6dU�¬/¦  «        “�d7 eR�d8d2�d�dd�dmdld�¬9¦  «        “�d: eR�d;d2�d�ddmdÈd°ddq�¬<¦	  «	        “�d= eR�d>d2�d�ddmdÈd°ddq�¬<¦	  «	        “�d? eR�d@d2�d�ddmdÊd°ddq�¬<¦	  «	        “�dA eR�dBd2�d�dd�dmdld�¬9¦  «        “¥i �dC eR�dDd2�d�ddmdÈd°ddq�¬<¦	  «	        “�dE eR�dFd2�d�ddmdÈd°ddq�¬<¦	  «	        “�dG eR�dHd2�d�d�dId�dmdld�¬J¦	  «	        “�dK eR�dLd2�d�d�dIdmdÈd°ddq�¬M¦
  «
        “�dN eR�dOd2�d�d�dIdmdÈd°ddq�¬M¦
  «
        “�dP eR�dQd2�d�d�dIdmdÊd°ddq�¬M¦
  «
        “�dR eR�dSd2�dTd»�dU�¬V¦  «        “�dW eR�dXd2�dTd\�dUd™¬½¦  «        “�dY eR�dZd2�d[d«d¼d¬¬½¦  «        “�d\ eRd2eed\d�d]dCdq�¬]¦  «        “�d^ eR�d_d2d\d�d]dž¬½¦  «        “�d` eRd2ee�dId\d�d]dž�¬a¦  «        “�db eR�dcd2¬3¦  «        “�dd eR�ded2¬3¦  «        “�df eR�dgd2¬3¦  «        “�dh eR�did2¬3¦  «        “�dj eR¦   «         “¥i �dk eR�dld2¬3¦  «        “�dm eR�dnd2¬3¦  «        “�do eR�dpd2¬3¦  «        “�dq eR�dT�dU�drd2�¬s¦  «        “�dt eR�du�dU�dvd2�¬s¦  «        “�dw eR�dx�dU�dyd2�¬s¦  «        “�dz eR�d{�d|�d}d2�¬s¦  «        “�d~ eR�d�d€�d�d2�¬s¦  «        “�d‚ eRd»dC�¬ƒ¦  «        “�d„ eRd»dC�¬ƒ¦  «        “�d… eRd»dC�¬ƒ¦  «        “�d† eRd2eedCd>d?¬D¦  «        “�d‡ eRd»dC�¬ƒ¦  «        “�dˆ eRd2�d‰�d|d?�¬Š¦  «        “�d‹ eRd2�d‰�d|d?�¬Š¦  «        “�dŒ eRd2�d�d�d‰�d|d?�¬�¦  «        “�dŽ eRd2�d�d�d‰�d|d?�¬�¦  «        “¥¦  «        ZSe5�d�d�e9f�d�„¦   «         ZTe5�d�d�e9f�d‘„¦   «         ZUe5�d�d�e9f�d’„¦   «         ZVe5�d�d�e9f�d“„¦   «         ZWe5�d�d�e9f�d”„¦   «         ZXe5�d�d�e9f�d•„¦   «         ZYe5�d�d�e9f�d–„¦   «         ZZe5�d�d�e9f�d—„¦   «         Z[e5�d�d�e9f�d˜„¦   «         Z\e5�d�d�e9f�d™„¦   «         Z]e5�d�d�e9f�dš„¦   «         Z^e5�d�d�e9f�d›„¦   «         Z_e5�d�d�e9f�dœ„¦   «         Z`e5�d�d�e9f�d�„¦   «         Zae5�d�d�e9f�dž„¦   «         Zbe5�d�d�e9f�dŸ„¦   «         Zce5�d�d�e9f�d „¦   «         Zde5�d�d�e9f�d¡„¦   «         Zee5�d�d�e9f�d¢„¦   «         Zfe5�d�d�e9f�d£„¦   «         Zge5�d�d�e9f�d¤„¦   «         Zhe5�d�d�e9f�d¥„¦   «         Zie5�d�d�e9f�d¦„¦   «         Zje5�d�d�e9f�d§„¦   «         Zke5�d�d�e9f�d¨„¦   «         Zle5�d�d�e9f�d©„¦   «         Zme5�d�d�e9f�dª„¦   «         Zne5�d�d�e9f�d«„¦   «         Zoe5�d�d�e9f�d¬„¦   «         Zpe5�d�d�e9f�d­„¦   «         Zqe5�d�d�e9f�d®„¦   «         Zre5�d�d�e9f�d¯„¦   «         Zse5�d�d�e9f�d°„¦   «         Zte5�d�d�e9f�d±„¦   «         Zue5�d�d�e9f�d²„¦   «         Zve5�d�d�e9f�d³„¦   «         Zwe5�d�d�e9f�d´„¦   «         Zxe5�d�d�e9f�dµ„¦   «         Zye5�d�d�e9f�d¶„¦   «         Zze5�d�d�e9f�d·„¦   «         Z{e5�d�d�e9f�d¸„¦   «         Z|e5�d�d�e9f�d¹„¦   «         Z}e5�d�d�e9f�dº„¦   «         Z~e5�d�d�e9f�d»„¦   «         Ze5�d�d�e9f�d¼„¦   «         Z€e5�d�d�e9f�d½„¦   «         Z�e5�d�d�e9f�d¾„¦   «         Z‚e5�d�d�e9f�d¿„¦   «         Zƒe5�d�d�e9f�dÀ„¦   «         Z„e5�d�d�e9f�dÁ„¦   «         Z…e5�d�d�e9f�dÂ„¦   «         Z†e5�d�d�e9f�dÃ„¦   «         Z‡e5�d�d�e9f�dÄ„¦   «         Zˆe5�d�d�e9f�dÅ„¦   «         Z‰e5�d�d�e9f�dÆ„¦   «         ZŠe5�d�d�e9f�dÇ„¦   «         Z‹e5�d�d�e9f�dÈ„¦   «         ZŒe5�d�d�e9f�dÉ„¦   «         Z�e5�d�d�e9f�dÊ„¦   «         ZŽe5�d�d�e9f�dË„¦   «         Z�e5�d�d�e9f�dÌ„¦   «         Z�e5�d�d�e9f�dÍ„¦   «         Z‘e5�d�d�e9f�dÎ„¦   «         Z’e5�d�d�e9f�dÏ„¦   «         Z“e5�d�d�e9f�dÐ„¦   «         Z”e5�d�d�e9f�dÑ„¦   «         Z•e5�d�d�e9f�dÒ„¦   «         Z–e5�d�d�e9f�dÓ„¦   «         Z—e5�d�d�e9f�dÔ„¦   «         Z˜e5�d�d�e9f�dÕ„¦   «         Z™e5�d�d�e9f�dÖ„¦   «         Zše5�d�d�e9f�d×„¦   «         Z›e5�d�d�e9f�dØ„¦   «         Zœe5�d�d�e9f�dÙ„¦   «         Z�e5�d�d�e9f�dÚ„¦   «         Zže5�d�d�e9f�dÛ„¦   «         ZŸe5�d�d�e9f�dÜ„¦   «         Z e5�d�d�e9f�dÝ„¦   «         Z¡e5�d�d�e9f�dÞ„¦   «         Z¢e5�d�d�e9f�dß„¦   «         Z£e5�d�d�e9f�dà„¦   «         Z¤e5�d�d�e9f�dá„¦   «         Z¥e5�d�d�e9f�dâ„¦   «         Z¦e5�d�d�e9f�dã„¦   «         Z§e5�d�d�e9f�dä„¦   «         Z¨e5�d�d�e9f�då„¦   «         Z©e5�d�d�e9f�dæ„¦   «         Zªe5�d�d�e9f�dç„¦   «         Z«e5�d�d�e9f�dè„¦   «         Z¬e5�d�d�e9f�dé„¦   «         Z­e5�d�d�e9f�dê„¦   «         Z®e5�d�d�e9f�dë„¦   «         Z¯e5�d�d�e9f�dì„¦   «         Z°e5�d�d�e9f�dí„¦   «         Z±e5�d�d�e9f�dî„¦   «         Z²e5�d�d�e9f�dï„¦   «         Z³e5�d�d�e9f�dð„¦   «         Z´e5�d�d�e9f�dñ„¦   «         Zµe5�d�d�e9f�dò„¦   «         Z¶e5�d�d�e9f�dó„¦   «         Z·e5�d�d�e9f�dô„¦   «         Z¸e5�d�d�e9f�dõ„¦   «         Z¹e5�d�d�e9f�dö„¦   «         Zºe5�d�d�e9f�d÷„¦   «         Z»e5�d�d�e9f�dø„¦   «         Z¼e5�d�d�e9f�dù„¦   «         Z½e5�d�d�e9f�dú„¦   «         Z¾e5�d�d�e9f�dû„¦   «         Z¿e5�d�d�e9f�dü„¦   «         ZÀe5�d�d�e9f�dý„¦   «         ZÁe5�d�d�e9f�dþ„¦   «         ZÂe5�d�d�e9f�dÿ„¦   «         ZÃe5�d�d�e9f�d „¦   «         ZÄe5�d�d�e9f�d„¦   «         ZÅe5�d�d�e9f�d„¦   «         ZÆ e6eÇi �ddæ“�ddé“�ddì“�ddî“�ddð“�ddò“�d	dô“�d
dö“�ddø“�ddË“�ddÎ“�ddÐ“�ddÒ“�ddÔ“�ddÖ“�ddÛ“�ddÝ“dßdã�d7�d:�d=�d?�dG�dK�dN�dP�d�d�d�d�dœ¥¦  «         dS (#  a   The EfficientNet Family in PyTorch

An implementation of EfficienNet that covers variety of related models with efficient architectures:

* EfficientNet-V2
  - `EfficientNetV2: Smaller Models and Faster Training` - https://arxiv.org/abs/2104.00298

* EfficientNet (B0-B8, L2 + Tensorflow pretrained AutoAug/RandAug/AdvProp/NoisyStudent weight ports)
  - EfficientNet: Rethinking Model Scaling for CNNs - https://arxiv.org/abs/1905.11946
  - CondConv: Conditionally Parameterized Convolutions for Efficient Inference - https://arxiv.org/abs/1904.04971
  - Adversarial Examples Improve Image Recognition - https://arxiv.org/abs/1911.09665
  - Self-training with Noisy Student improves ImageNet classification - https://arxiv.org/abs/1911.04252

* MixNet (Small, Medium, and Large)
  - MixConv: Mixed Depthwise Convolutional Kernels - https://arxiv.org/abs/1907.09595

* MNasNet B1, A1 (SE), Small
  - MnasNet: Platform-Aware Neural Architecture Search for Mobile - https://arxiv.org/abs/1807.11626

* FBNet-C
  - FBNet: Hardware-Aware Efficient ConvNet Design via Differentiable NAS - https://arxiv.org/abs/1812.03443

* Single-Path NAS Pixel1
  - Single-Path NAS: Designing Hardware-Efficient ConvNets - https://arxiv.org/abs/1904.02877

* TinyNet
    - Model Rubik's Cube: Twisting Resolution, Depth and Width for TinyNets - https://arxiv.org/abs/2010.14819
    - Definitions & weights borrowed from https://github.com/huawei-noah/CV-Backbones/tree/master/tinynet_pytorch

* And likely more...

The majority of the above models (EfficientNet*, MixNet, MnasNet) and original weights were made available
by Mingxing Tan, Quoc Le, and other members of their Google Brain team. Thanks for consistently releasing
the models and weights open source!

Hacked together by / Copyright 2019, Ross Wightman
é    )Úpartial)ÚCallableÚDictÚListÚOptionalÚTupleÚUnionN)ÚIMAGENET_DEFAULT_MEANÚIMAGENET_DEFAULT_STDÚIMAGENET_INCEPTION_MEANÚIMAGENET_INCEPTION_STD)Úcreate_conv2dÚcreate_classifierÚget_norm_act_layerÚ	LayerTypeÚGroupNormActÚLayerNormAct2dÚEvoNorm2dS0Úget_device_dtypeé   )Úbuild_model_with_cfgÚpretrained_cfg_for_features)ÚSqueezeExcite)Ú	BlockArgsÚEfficientNetBuilderÚdecode_arch_defÚefficientnet_init_weightsÚround_channelsÚresolve_bn_argsÚresolve_act_layerÚBN_EPS_TF_DEFAULT)ÚFeatureInfoÚFeatureHooksÚfeature_take_indices)Úcheckpoint_seqÚ
checkpoint)Úgenerate_default_cfgsÚregister_modelÚregister_model_deprecationsÚEfficientNetÚEfficientNetFeaturesc            %       ó2  ‡ — e Zd ZdZdddddddddddded	d	d
ddfdededededededededede	e
         de	e
         de	e
         de	e
         dededededdf$ˆ fd„Zdej        fd„Zej        j        d7dedeeeeef         f         fd „¦   «         Zej        j        d8d"eddfd#„¦   «         Zej        j        dej        fd$„¦   «         Zd9dededdfd%„Z	 	 	 	 	 	 d:d'ej        d(e	eeee         f                  d)ed*ed+ed,ed-edeeej                 eej        eej                 f         f         fd.„Z	 	 	 	 d;d(eeee         f         d0ed1ed-edee         f
d2„Zd'ej        dej        fd3„Z d7d'ej        d4edej        fd5„Z!d'ej        dej        fd6„Z"ˆ xZ#S )<r*   a’  EfficientNet model architecture.

    A flexible and performant PyTorch implementation of efficient network architectures, including:
      * EfficientNet-V2 Small, Medium, Large, XL & B0-B3
      * EfficientNet B0-B8, L2
      * EfficientNet-EdgeTPU
      * EfficientNet-CondConv
      * MixNet S, M, L, XL
      * MnasNet A1, B1, and small
      * MobileNet-V2
      * FBNet C
      * Single-Path NAS Pixel1
      * TinyNet

    References:
      - EfficientNet: https://arxiv.org/abs/1905.11946
      - EfficientNetV2: https://arxiv.org/abs/2104.00298
      - MixNet: https://arxiv.org/abs/1907.09595
      - MnasNet: https://arxiv.org/abs/1807.11626
    éè  é   é   é    FÚ Nç        ÚavgÚ
block_argsÚnum_classesÚnum_featuresÚin_chansÚ	stem_sizeÚstem_kernel_sizeÚfix_stemÚoutput_strideÚpad_typeÚ	act_layerÚ
norm_layerÚaa_layerÚse_layerÚround_chs_fnÚ	drop_rateÚdrop_path_rateÚglobal_poolÚreturnc                 óp  •— t          ¦   «                              ¦   «          ||dœ}|
pt          j        }
|pt          j        }t          ||
¦  «        }|pt          }|| _        || _        || _	        d| _
        |s ||¦  «        }t          |||fd|	dœ|¤Ž| _         ||fddi|¤Ž| _        t          d||	||
||||dœ|¤Ž}t          j         |||¦  «        Ž | _        |j        | _        d„ | j        D ¦   «         | _        |j        }|d	k    r5t          ||d
fd|	i|¤Ž| _         ||fddi|¤Ž| _        |x| _        | _        n>t          j        ¦   «         | _        t          j        ¦   «         | _        |x| _        | _        t5          | j        | j        fd|i|¤Ž\  | _        | _        t;          | ¦  «         dS )aÊ  Initialize EfficientNet model.

        Args:
            block_args: Arguments for building blocks.
            num_classes: Number of classifier classes.
            num_features: Number of features for penultimate layer.
            in_chans: Number of input channels.
            stem_size: Number of output channels in stem.
            stem_kernel_size: Kernel size for stem convolution.
            fix_stem: If True, don't scale stem channels.
            output_stride: Output stride of network.
            pad_type: Padding type.
            act_layer: Activation layer class.
            norm_layer: Normalization layer class.
            aa_layer: Anti-aliasing layer class.
            se_layer: Squeeze-and-excitation layer class.
            round_chs_fn: Channel rounding function.
            drop_rate: Dropout rate for classifier.
            drop_path_rate: Drop path rate for stochastic depth.
            global_pool: Global pooling type.
        ©ÚdeviceÚdtypeFé   ©ÚstrideÚpaddingÚinplaceT)r;   r<   rA   r=   r>   r?   r@   rC   c                 ó   — g | ]
}|d          ‘ŒS )Ústage© ©Ú.0Úfs     úh/root/private/project/pallet_detection_v2/.venv/lib/python3.11/site-packages/timm/models/efficientnet.pyú
<listcomp>z)EfficientNet.__init__.<locals>.<listcomp>›   s   € ÐAÐAÐA¨!˜1˜Wœ:ÐAÐAÐAó    r   r   rM   Ú	pool_typeNrQ   )ÚsuperÚ__init__ÚnnÚReLUÚBatchNorm2dr   r   r5   r7   rB   Úgrad_checkpointingr   Ú	conv_stemÚbn1r   Ú
SequentialÚblocksÚfeaturesÚfeature_infoÚ
stage_endsÚin_chsÚ	conv_headÚbn2r6   Úhead_hidden_sizeÚIdentityr   rD   Ú
classifierr   )Úselfr4   r5   r6   r7   r8   r9   r:   r;   r<   r=   r>   r?   r@   rA   rB   rC   rD   rH   rI   ÚddÚnorm_act_layerÚbuilderÚhead_chsÚ	__class__s                           €rU   rZ   zEfficientNet.__init__Q   s5  ø€ õV 	‰Œ×ÒÑÔÐØ¨Ð/Ð/ˆØÐ(¥¤ˆ	ØÐ1¥2¤>ˆ
Ý+¨J¸	ÑBÔBˆØÐ,�}ˆØ&ˆÔØ ˆŒØ"ˆŒØ"'ˆÔð ð 	0Ø$˜ YÑ/Ô/ˆIÝ& x°Ð<LÐoÐUVÐ`hÐoÐoÐlnÐoÐoˆŒØ!�> )Ð@Ð@°TÐ@¸RÐ@Ð@ˆŒõ &ð 

Ø'ØØ%ØØ!ØØØ)ð

ð 

ð ð

ð 

ˆõ ”m W W¨Y¸
Ñ%CÔ%CÐDˆŒØ#Ô,ˆÔØAÐA¨tÔ/@ÐAÑAÔAˆŒØ”>ˆð ˜!ÒÐÝ*¨8°\À1Ð]Ð]ÈhÐ]ÐZ\Ð]Ð]ˆDŒNØ%�~ lÐGÐG¸DÐGÀBÐGÐGˆDŒHØ8DÐDˆDÔ Ô 5Ð 5åœ[™]œ]ˆDŒNÝ”{‘}”}ˆDŒHØ8@Ð@ˆDÔ Ô 5å,=ØÔØÔð-
ð -
ð "ð-
ð ð	-
ð -
Ñ)ˆÔ˜$œ/õ 	" $Ñ'Ô'Ð'Ð'Ð'rW   c                 ó"  — | j         | j        g}|                     | j        ¦  «         |                     | j        | j        | j        g¦  «         |                     t          j        | j	        ¦  «        | j
        g¦  «         t          j        |Ž S )z3Convert model to sequential for feature extraction.)r_   r`   Úextendrb   rg   rh   rD   r[   ÚDropoutrB   rk   ra   )rl   Úlayerss     rU   Úas_sequentialzEfficientNet.as_sequential±   sv   € à”. $¤(Ð+ˆØ�Š�d”kÑ"Ô"Ð"Ø�Š�t”~ t¤x°Ô1AÐBÑCÔCÐCØ�Š•r”z $¤.Ñ1Ô1°4´?ÐCÑDÔDÐDÝŒ}˜fÐ%Ð%rW   Úcoarsec                 ó4   — t          d|rdnddfdg¬¦  «        S )zÊCreate regex patterns for parameter groups.

        Args:
            coarse: Use coarse (stage-level) grouping.

        Returns:
            Dictionary mapping group names to regex patterns.
        z^conv_stem|bn1z^blocks\.(\d+)z^blocks\.(\d+)\.(\d+)N)zconv_head|bn2)iŸ† )Ústemrb   )Údict)rl   rw   s     rU   Úgroup_matcherzEfficientNet.group_matcher¹   s:   € õ Ø"à&,ÐJÐ"Ð"Ð2JÈDÐQØ,ðð
ñ 
ô 
ð 	
rW   TÚenablec                 ó   — || _         dS ©zEnable or disable gradient checkpointing.

        Args:
            enable: Whether to enable gradient checkpointing.
        N©r^   ©rl   r|   s     rU   Úset_grad_checkpointingz#EfficientNet.set_grad_checkpointingË   ó   € ð #)ˆÔÐÐrW   c                 ó   — | j         S )zGet the classifier module.)rk   )rl   s    rU   Úget_classifierzEfficientNet.get_classifierÔ   s   € ð ŒÐrW   c                 óü   — t          | ¦  «        }|| _        t          | j        | j        fd|i|¤Ž\  | _        | _        | j                             | j        ¦  «         | j                             | j        ¦  «         dS )zŸReset the classifier head.

        Args:
            num_classes: Number of classes for new classifier.
            global_pool: Global pooling type.
        rX   N)r   r5   r   r6   rD   rk   ÚtrainÚtraining)rl   r5   rD   rm   s       rU   Úreset_classifierzEfficientNet.reset_classifierÙ   sŽ   € õ ˜dÑ#Ô#ˆØ&ˆÔÝ,=ØÔ˜tÔ/ð-Nð -NØ;Fð-NØJLð-Nð -NÑ)ˆÔ˜$œ/àÔ×Ò˜tœ}Ñ-Ô-Ð-ØŒ×Ò˜dœmÑ,Ô,Ð,Ð,Ð,rW   ÚNCHWÚxÚindicesÚnormÚ
stop_earlyÚ
output_fmtÚintermediates_onlyÚextra_blocksc                 óZ  ‡ — |dv s
J d¦   «         ‚g }|r)t          t          ‰ j        ¦  «        dz   |¦  «        \  }	}
n@t          t          ‰ j        ¦  «        |¦  «        \  }	}
ˆ fd„|	D ¦   «         }	‰ j        |
         }
d}‰                      |¦  «        }‰                      |¦  «        }||	v r|                     |¦  «         t          j         	                    ¦   «         s|s‰ j        }n‰ j        d|
…         }t          |d¬¦  «        D ]_\  }}‰ j        r/t          j         	                    ¦   «         st          ||¦  «        }n ||¦  «        }||	v r|                     |¦  «         Œ`|r|S |‰ j        d         k    r*‰                      |¦  «        }‰                      |¦  «        }||fS )	a¾  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.
            extra_blocks: Include outputs of all blocks and head conv in output, does not align with feature_info.

        Returns:
            List of intermediate features or tuple of (final features, intermediates).
        )r‰   zOutput shape must be NCHW.r   c                 ó*   •— g | ]}‰j         |         ‘ŒS rQ   )re   )rS   Úirl   s     €rU   rV   z6EfficientNet.forward_intermediates.<locals>.<listcomp>  s    ø€ ÐEÐEÐE°1˜DœO¨AÔ.ÐEÐEÐErW   r   N)Ústartéÿÿÿÿ)r$   Úlenrb   re   r_   r`   ÚappendÚtorchÚjitÚis_scriptingÚ	enumerater^   r%   rg   rh   )rl   rŠ   r‹   rŒ   r�   rŽ   r�   r�   ÚintermediatesÚtake_indicesÚ	max_indexÚfeat_idxrb   Úblks   `             rU   Úforward_intermediatesz"EfficientNet.forward_intermediatesç   sÙ  ø€ ð0 ˜YÐ&Ð&Ð&Ð(DÑ&Ô&Ð&ØˆØð 	3Ý&:½3¸t¼{Ñ;KÔ;KÈaÑ;OÐQXÑ&YÔ&YÑ#ˆL˜)˜)å&:½3¸t¼Ñ;OÔ;OÐQXÑ&YÔ&YÑ#ˆL˜)ØEÐEÐEÐE¸ÐEÑEÔEˆLØœ¨	Ô2ˆIàˆØ�NŠN˜1ÑÔˆØ�HŠH�Q‰KŒKˆØ�|Ð#Ð#Ø× Ò  Ñ#Ô#Ð#åŒ9×!Ò!Ñ#Ô#ð 	-¨:ð 	-Ø”[ˆFˆFà”[  ) Ô,ˆFÝ& v°QÐ7Ñ7Ô7ð 	(ð 	(‰MˆH�cØÔ&ð ­u¬y×/EÒ/EÑ/GÔ/Gð Ý" 3¨Ñ*Ô*��à�C˜‘F”F�Ø˜<Ð'Ð'Ø×$Ò$ QÑ'Ô'Ð'øàð 	!Ø Ð à�t” rÔ*Ò*Ð*Ø—’˜qÑ!Ô!ˆAØ—’˜‘”ˆAà�-ÐÐrW   r   Ú
prune_normÚ
prune_headc                 ó¬  — |r)t          t          | j        ¦  «        dz   |¦  «        \  }}n2t          t          | j        ¦  «        |¦  «        \  }}| j        |         }| j        d|…         | _        |s|t          | j        ¦  «        k     r0t	          j        ¦   «         | _        t	          j        ¦   «         | _        |r|                      dd¦  «         |S )a€  Prune layers not required for specified intermediates.

        Args:
            indices: Indices of intermediate layers to keep.
            prune_norm: Whether to prune normalization layers.
            prune_head: Whether to prune the classifier head.
            extra_blocks: Include all blocks in indexing.

        Returns:
            List of indices that were kept.
        r   Nr   r1   )	r$   r–   rb   re   r[   rj   rg   rh   rˆ   )rl   r‹   r¢   r£   r�   r�   rž   s          rU   Úprune_intermediate_layersz&EfficientNet.prune_intermediate_layers#  sÉ   € ð$ ð 	3Ý&:½3¸t¼{Ñ;KÔ;KÈaÑ;OÐQXÑ&YÔ&YÑ#ˆL˜)˜)å&:½3¸t¼Ñ;OÔ;OÐQXÑ&YÔ&YÑ#ˆL˜)Øœ¨	Ô2ˆIØ”k * 9 *Ô-ˆŒØð 	%˜¥S¨¬Ñ%5Ô%5Ò5Ð5Ýœ[™]œ]ˆDŒNÝ”{‘}”}ˆDŒHØð 	)Ø×!Ò! ! RÑ(Ô(Ð(ØÐrW   c                 óR  — |                       |¦  «        }|                      |¦  «        }| j        r6t          j                             ¦   «         st          | j        |d¬¦  «        }n|                      |¦  «        }|                      |¦  «        }|  	                    |¦  «        }|S )z/Forward pass through feature extraction layers.T)Úflatten)
r_   r`   r^   r˜   r™   rš   r%   rb   rg   rh   ©rl   rŠ   s     rU   Úforward_featureszEfficientNet.forward_featuresB  sŽ   € à�NŠN˜1ÑÔˆØ�HŠH�Q‰KŒKˆØÔ"ð 	­5¬9×+AÒ+AÑ+CÔ+Cð 	Ý˜tœ{¨A°tÐ<Ñ<Ô<ˆAˆAà—’˜A‘”ˆAØ�NŠN˜1ÑÔˆØ�HŠH�Q‰KŒKˆØˆrW   Ú
pre_logitsc                 ó¶   — |                       |¦  «        }| j        dk    r!t          j        || j        | j        ¬¦  «        }|r|n|                      |¦  «        S )zÊForward pass through classifier head.

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

        Returns:
            Output tensor.
        r2   )Úpr‡   )rD   rB   ÚFÚdropoutr‡   rk   )rl   rŠ   rª   s      rU   Úforward_headzEfficientNet.forward_headN  sY   € ð ×Ò˜QÑÔˆØŒ>˜BÒÐÝ”	˜!˜tœ~¸¼ÐFÑFÔFˆAØÐ6ˆqˆq D§O¢O°AÑ$6Ô$6Ð6rW   c                 óZ   — |                       |¦  «        }|                      |¦  «        }|S )zForward pass.)r©   r¯   r¨   s     rU   ÚforwardzEfficientNet.forward]  s-   € à×!Ò! !Ñ$Ô$ˆØ×Ò˜aÑ Ô ˆØˆrW   ©F©T)r3   )NFFr‰   FF)r   FTF)$Ú__name__Ú
__module__Ú__qualname__Ú__doc__r   r   ÚintÚboolÚstrr   r   r   ÚfloatrZ   r[   ra   rv   r˜   r™   Úignorer   r	   r   r{   r�   ÚModuler„   rˆ   ÚTensorr   r¡   r¥   r©   r¯   r±   Ú__classcell__©rq   s   @rU   r*   r*   ;   sý  ø€ € € € € ðð ð0  $Ø $ØØØ$%Ø"Ø!#ØØ-1Ø.2Ø,0Ø,0Ø%3Ø!Ø$&Ø$ØØð)^(ð ^(à!ð^(ð ð^(ð ð	^(ð
 ð^(ð ð^(ð "ð^(ð ð^(ð ð^(ð ð^(ð   	Ô*ð^(ð ! Ô+ð^(ð ˜yÔ)ð^(ð ˜yÔ)ð^(ð #ð^(ð  ð!^(ð" "ð#^(ð$ ð%^(ð* 
ð+^(ð ^(ð ^(ð ^(ð ^(ð ^(ð@&˜rœ}ð &ð &ð &ð &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-ð -¨Cð -¸cð -Èdð -ð -ð -ð -ð" 8<ØØ$Ø$Ø',Ø!&ð: ð : àŒ|ð: ð ˜e C¨¨c¬ NÔ3Ô4ð: ð ð	: ð
 ð: ð ð: ð !%ð: ð ð: ð 
ˆt�E”LÔ! 5¨¬°t¸E¼LÔ7IÐ)IÔ#JÐJÔ	Kð: ð : ð : ð : ð| ./Ø$Ø#Ø!&ðð à˜3  S¤	˜>Ô*ðð ðð ð	ð
 ðð 
ˆcŒðð ð ð ð>
 %¤,ð 
°5´<ð 
ð 
ð 
ð 
ð7ð 7˜eœlð 7¸ð 7ÈÌð 7ð 7ð 7ð 7ð˜œð ¨%¬,ð ð ð ð ð ð ð ð rW   c            !       ó   ‡ — e Zd ZdZdddddddddddded	d	ddfd
edeedf         dedededede	dedede
e         de
e         de
e         de
e         dededef ˆ fd„Zej        j        d!de	ddfd„¦   «         Zdeej                 fd „Zˆ xZS )"r+   z¯ EfficientNet Feature Extractor

    A work-in-progress feature extraction module for EfficientNet, to use as a backbone for segmentation
    and object detection models.
    )r   r   rJ   r/   é   Ú
bottleneckr/   r0   Fr1   Nr2   r4   Úout_indices.Úfeature_locationr7   r8   r9   r:   r;   r<   r=   r>   r?   r@   rA   rB   rC   c                 óø  •— t          ¦   «                              ¦   «          ||dœ}|
pt          j        }
|pt          j        }t          ||
¦  «        }|pt          }|| _        || _        d| _	        |s ||¦  «        }t          |||fd|	dœ|¤Ž| _         ||fddi|¤Ž| _        t          d||	||
|||||dœ	|¤Ž}t          j         |||¦  «        Ž | _        t!          |j        |¦  «        | _        d„ | j                             ¦   «         D ¦   «         | _        t+          | ¦  «         d | _        |d	k    rD| j                             d
¬¦  «        }t/          ||                      ¦   «         ¦  «        | _        d S d S )NrG   FrJ   rK   rN   T)	r;   r<   rA   r=   r>   r?   r@   rC   rÅ   c                 ó,   — i | ]}|d          |d         “ŒS )rP   ÚindexrQ   rR   s     rU   ú
<dictcomp>z1EfficientNetFeatures.__init__.<locals>.<dictcomp>Ÿ  s"   € Ð]Ð]Ð]¸!˜q œz¨1¨W¬:Ð]Ð]Ð]rW   rÃ   )ÚmoduleÚ	hook_type)ÚkeysrQ   )rY   rZ   r[   r\   r]   r   r   r7   rB   r^   r   r_   r`   r   ra   rb   r"   rc   rd   Ú	get_dictsÚ_stage_out_idxr   Úfeature_hooksr#   Únamed_modules)rl   r4   rÄ   rÅ   r7   r8   r9   r:   r;   r<   r=   r>   r?   r@   rA   rB   rC   rH   rI   rm   rn   ro   Úhooksrq   s                          €rU   rZ   zEfficientNetFeatures.__init__k  sË  ø€ õ* 	‰Œ×ÒÑÔÐØ¨Ð/Ð/ˆØÐ(¥¤ˆ	ØÐ1¥2¤>ˆ
Ý+¨J¸	ÑBÔBˆØÐ,�}ˆØ ˆŒØ"ˆŒØ"'ˆÔð ð 	0Ø$˜ YÑ/Ô/ˆIÝ& x°Ð<LÐoÐUVÐ`hÐoÐoÐlnÐoÐoˆŒØ!�> )Ð@Ð@°TÐ@¸RÐ@Ð@ˆŒõ &ð 
Ø'ØØ%ØØ!ØØØ)Ø-ð
ð 
ð ð
ð 
ˆõ ”m W W¨Y¸
Ñ%CÔ%CÐDˆŒÝ'¨Ô(8¸+ÑFÔFˆÔØ]Ð]¸tÔ?P×?ZÒ?ZÑ?\Ô?\Ð]Ñ]Ô]ˆÔå! $Ñ'Ô'Ð'ð "ˆÔØ˜|Ò+Ð+ØÔ%×/Ò/Ð5LÐ/ÑMÔMˆEÝ!-¨e°T×5GÒ5GÑ5IÔ5IÑ!JÔ!JˆDÔÐÐð ,Ð+rW   Tr|   rE   c                 ó   — || _         dS r~   r   r€   s     rU   r�   z+EfficientNetFeatures.set_grad_checkpointing©  r‚   rW   c                 óJ  — |                       |¦  «        }|                      |¦  «        }| j        €žg }d| j        v r|                     |¦  «         t          | j        ¦  «        D ]g\  }}| j        r/t          j	         
                    ¦   «         st          ||¦  «        }n ||¦  «        }|dz   | j        v r|                     |¦  «         Œh|S |                      |¦  «         | j                             |j        ¦  «        }t          |                     ¦   «         ¦  «        S )Nr   r   )r_   r`   rÏ   rÎ   r—   r›   rb   r^   r˜   r™   rš   r&   Ú
get_outputrH   ÚlistÚvalues)rl   rŠ   rc   r“   ÚbÚouts         rU   r±   zEfficientNetFeatures.forward²  s  € Ø�NŠN˜1ÑÔˆØ�HŠH�Q‰KŒKˆØÔÐ%ØˆHØ�DÔ'Ð'Ð'Ø—’ Ñ"Ô"Ð"Ý! $¤+Ñ.Ô.ð 'ð '‘��1ØÔ*ð µ5´9×3IÒ3IÑ3KÔ3Kð Ý" 1 aÑ(Ô(�A�Aà˜˜!™œ�AØ�q‘5˜DÔ/Ð/Ð/Ø—O’O AÑ&Ô&Ð&øØˆOà�KŠK˜‰NŒNˆNØÔ$×/Ò/°´Ñ9Ô9ˆCÝ˜Ÿ
š
™œÑ%Ô%Ð%rW   r³   )r´   rµ   r¶   r·   r   r   r   r¸   rº   r¹   r   r   r   r»   rZ   r˜   r™   r¼   r�   r   r¾   r±   r¿   rÀ   s   @rU   r+   r+   d  s®  ø€ € € € € ðð ð ,;Ø$0ØØØ$%Ø"Ø!#ØØ-1Ø.2Ø,0Ø,0Ø%3Ø!Ø$&ØØð'<Kð <Kà!ð<Kð ˜s C˜xœð<Kð "ð	<Kð
 ð<Kð ð<Kð "ð<Kð ð<Kð ð<Kð ð<Kð   	Ô*ð<Kð ! Ô+ð<Kð ˜yÔ)ð<Kð ˜yÔ)ð<Kð #ð<Kð  ð!<Kð" "ð#<Kð <Kð <Kð <Kð <Kð <Kð| „YÔð)ð )¨Tð )¸Tð )ð )ð )ñ Ôð)ð&˜D ¤Ô.ð &ð &ð &ð &ð &ð &ð &ð &rW   Fc                 ó&  — d}t           }d }|                     dd¦  «        rd|v sd|v rd}nd}t          }d}|                     d	d
¦  «        }t          || |f|dk    |o|dk    |dœ|¤Ž}|dk    r t	          |j        ¦  «        x|_        |_        |S )Nr1   Úfeatures_onlyFÚfeature_cfgÚfeature_clsÚcfg)r5   r6   Ú	head_convrD   ÚclsÚpretrained_strictT)rÚ   rà   Úkwargs_filter)r*   Úpopr+   r   r   Úpretrained_cfgÚdefault_cfg)ÚvariantÚ
pretrainedÚkwargsÚfeatures_modeÚ	model_clsrá   rà   Úmodels           rU   Ú_create_effnetrë   Ç  sÚ   € Ø€MÝ€IØ€MØ‡z‚z�/ 5Ñ)Ô)ð "Ø˜FÐ"Ð" m°vÐ&=Ð&=Ø!ˆMˆMàWˆMÝ,ˆIØ!ˆMØŸ
š
Ð#6¸Ñ=Ô=Ðå ØØØðð $ uÒ,Ø+ÐF°ÀÒ0FØ#ðð ð ðð €Eð ˜ÒÐÝ3NÈuÔOcÑ3dÔ3dÐdˆÔ˜uÔ0Ø€LrW   ç      ð?c                 ó  — dgdgdgdgdgdgdgg}t          dt          |¦  «        dt          t          |¬	¦  «        |                     d
d¦  «        p#t          t
          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )z÷Creates a mnasnet-a1 model.

    Ref impl: https://github.com/tensorflow/tpu/tree/master/models/official/mnasnet
    Paper: https://arxiv.org/pdf/1807.11626.pdf.

    Args:
      channel_multiplier: multiplier to number of channels per layer.
    Úds_r1_k3_s1_e1_c16_noskipÚir_r2_k3_s2_e6_c24zir_r3_k5_s2_e3_c40_se0.25Úir_r4_k3_s2_e6_c80zir_r2_k3_s1_e6_c112_se0.25zir_r3_k5_s2_e6_c160_se0.25Úir_r1_k3_s1_e6_c320r0   ©Ú
multiplierr>   N©r4   r8   rA   r>   rQ   ©	rz   r   r   r   râ   r[   r]   r   rë   ©rå   Úchannel_multiplierræ   rç   Úarch_defÚmodel_kwargsrê   s          rU   Ú_gen_mnasnet_a1rú   â  sÈ   € ð 
%Ð%à	Ðà	$Ð%à	Ðà	%Ð&à	%Ð&à	Ðð€Hõ  ð Ý" 8Ñ,Ô,ØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gð	ð ð
 ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó  — dgdgdgdgdgdgdgg}t          dt          |¦  «        dt          t          |¬	¦  «        |                     d
d¦  «        p#t          t
          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )ú÷Creates a mnasnet-b1 model.

    Ref impl: https://github.com/tensorflow/tpu/tree/master/models/official/mnasnet
    Paper: https://arxiv.org/pdf/1807.11626.pdf.

    Args:
      channel_multiplier: multiplier to number of channels per layer.
    Úds_r1_k3_s1_c16_noskipÚir_r3_k3_s2_e3_c24Úir_r3_k5_s2_e3_c40Úir_r3_k5_s2_e6_c80Úir_r2_k3_s1_e6_c96Úir_r4_k5_s2_e6_c192Úir_r1_k3_s1_e6_c320_noskipr0   rò   r>   Nrô   rQ   rõ   rö   s          rU   Ú_gen_mnasnet_b1r    sÈ   € ð 
"Ð"à	Ðà	Ðà	Ðà	Ðà	Ðà	%Ð&ð€Hõ  ð Ý" 8Ñ,Ô,ØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gð	ð ð
 ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó  — dgdgdgdgdgdgdgg}t          dt          |¦  «        dt          t          |¬	¦  «        |                     d
d¦  «        p#t          t
          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )rü   Úds_r1_k3_s1_c8Úir_r1_k3_s2_e3_c16Úir_r2_k3_s2_e6_c16zir_r4_k5_s2_e6_c32_se0.25zir_r3_k3_s1_e6_c32_se0.25zir_r3_k5_s2_e6_c88_se0.25Úir_r1_k3_s1_e6_c144é   rò   r>   Nrô   rQ   rõ   rö   s          rU   Ú_gen_mnasnet_smallr  *  sÈ   € ð 
ÐØ	ÐØ	ÐØ	$Ð%Ø	$Ð%Ø	$Ð%Ø	Ðð€Hõ ð Ý" 8Ñ,Ô,ØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gð	ð ð
 ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó~  — dgdgdgdgdgg}t          t          |¬¦  «        }	|r|rdnt          d |	d¦  «        ¦  «        nd}
t          dt	          ||||¬	¦  «        |
d
||	|                     dd¦  «        p#t          t          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}t          | |fi |¤Ž}|S )z�
    Ref impl: https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v2.py
    Paper: https://arxiv.org/abs/1801.04381
    Údsa_r1_k3_s1_c64Údsa_r2_k3_s2_c128Údsa_r2_k3_s2_c256Údsa_r6_k3_s2_c512Údsa_r2_k3_s2_c1024rò   i   r   ©Údepth_multiplierÚfix_first_lastÚ
group_sizer0   r>   NÚrelu6©r4   r6   r8   r:   rA   r>   r=   rQ   )r   r   Úmaxrz   r   râ   r[   r]   r   r    rë   )rå   r÷   r  r  Úfix_stem_headrÞ   ræ   rç   rø   rA   Úhead_featuresrù   rê   s                rU   Ú_gen_mobilenet_v1r  G  s  € ð 
ÐØ	ÐØ	ÐØ	ÐØ	Ðð€Hõ �>Ð6HÐIÑIÔI€LØR[Ðb˜]ÐM�T�Tµ°D¸,¸,ÀtÑ:LÔ:LÑ0MÔ0MÐ0MÐab€MÝð Ý"ØØ-Ø(Ø!ð	
ñ 
ô 
ð #ØØØ!Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨GÑ4Ô4ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 óz  — dgdgdgdgdgdgdgg}t          t          |¬¦  «        }t          dt          ||||¬	¦  «        |rd
nt	          d
 |d
¦  «        ¦  «        d|||                     dd¦  «        p#t          t          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}	t          | |fi |	¤Ž}
|
S )z» Generate MobileNet-V2 network
    Ref impl: https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v2.py
    Paper: https://arxiv.org/abs/1801.04381
    Úds_r1_k3_s1_c16rï   Úir_r3_k3_s2_e6_c32Úir_r4_k3_s2_e6_c64Úir_r3_k3_s1_e6_c96Úir_r3_k3_s2_e6_c160rñ   rò   r  r.   r0   r>   Nr  r  rQ   )r   r   rz   r   r  râ   r[   r]   r   r    rë   )rå   r÷   r  r  r  ræ   rç   rø   rA   rù   rê   s              rU   Ú_gen_mobilenet_v2r"  k  s  € ð 
ÐØ	ÐØ	ÐØ	ÐØ	ÐØ	ÐØ	Ðð€Hõ �>Ð6HÐIÑIÔI€LÝð Ý"ØØ-Ø(Ø!ð	
ñ 
ô 
ð +ÐM�T�Tµ°D¸,¸,ÀtÑ:LÔ:LÑ0MÔ0MØØØ!Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨GÑ4Ô4ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó   — dgddgg d¢g d¢ddgdgd	gg}t          dt          |¦  «        d
dt          t          |¬¦  «        |                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )ai   FBNet-C

        Paper: https://arxiv.org/abs/1812.03443
        Ref Impl: https://github.com/facebookresearch/maskrcnn-benchmark/blob/master/maskrcnn_benchmark/modeling/backbone/fbnet_modeldef.py

        NOTE: the impl above does not relate to the 'C' variant here, that was derived from paper,
        it was used to confirm some building block details
    Úir_r1_k3_s1_e1_c16Úir_r1_k3_s2_e6_c24Úir_r2_k3_s1_e1_c24)Úir_r1_k5_s2_e6_c32Úir_r1_k5_s1_e3_c32Úir_r1_k5_s1_e6_c32Úir_r1_k3_s1_e6_c32)Úir_r1_k5_s2_e6_c64Úir_r1_k5_s1_e3_c64Úir_r2_k5_s1_e6_c64Úir_r3_k5_s1_e6_c112Úir_r1_k5_s1_e3_c112Úir_r4_k5_s2_e6_c184Úir_r1_k3_s1_e6_c352é   iÀ  rò   r>   N)r4   r8   r6   rA   r>   rQ   rõ   rö   s          rU   Ú_gen_fbnetcr3  �  s×   € ð 
ÐØ	Ð3Ð4Ø`Ð`Ð`ØJÐJÐJØ	Ð 5Ð6Ø	ÐØ	Ðð€Hõ ð Ý" 8Ñ,Ô,ØØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó  — dgdgddgddgddgd	gd
gg}t          dt          |¦  «        dt          t          |¬¦  «        |                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )zËCreates the Single-Path NAS model from search targeted for Pixel1 phone.

    Paper: https://arxiv.org/abs/1904.02877

    Args:
      channel_multiplier: multiplier to number of channels per layer.
    rý   rþ   Úir_r1_k5_s2_e6_c40Úir_r3_k3_s1_e3_c40Úir_r1_k5_s2_e6_c80Úir_r3_k3_s1_e3_c80Úir_r1_k5_s1_e6_c96Úir_r3_k5_s1_e3_c96r  r  r0   rò   r>   Nrô   rQ   rõ   rö   s          rU   Ú_gen_spnasnetr;  ®  sÑ   € ð 
"Ð"à	Ðà	Ð3Ð4à	Ð3Ð4à	Ð3Ð4à	Ðà	%Ð&ð€Hõ  ð Ý" 8Ñ,Ô,ØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gð	ð ð
 ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   r
  c                 óT  — dgdgdgdgdgdgdgg}t          t          ||¬¦  «        }t          dt          |||¬	¦  «         |d
¦  «        d|t	          |d¦  «        |                     dd¦  «        p#t          t          j        fi t          |¦  «        ¤Ždœ|¤Ž}	t          | |fi |	¤Ž}
|
S )ax  Creates an EfficientNet model.

    Ref impl: https://github.com/tensorflow/tpu/blob/master/models/official/efficientnet/efficientnet_model.py
    Paper: https://arxiv.org/abs/1905.11946

    EfficientNet params
    name: (channel_multiplier, depth_multiplier, resolution, dropout_rate)
    'efficientnet-b0': (1.0, 1.0, 224, 0.2),
    'efficientnet-b1': (1.0, 1.1, 240, 0.2),
    'efficientnet-b2': (1.1, 1.2, 260, 0.3),
    'efficientnet-b3': (1.2, 1.4, 300, 0.3),
    'efficientnet-b4': (1.4, 1.8, 380, 0.4),
    'efficientnet-b5': (1.6, 2.2, 456, 0.4),
    'efficientnet-b6': (1.8, 2.6, 528, 0.5),
    'efficientnet-b7': (2.0, 3.1, 600, 0.5),
    'efficientnet-b8': (2.2, 3.6, 672, 0.5),
    'efficientnet-l2': (4.3, 5.3, 800, 0.5),

    Args:
      channel_multiplier: multiplier to number of channels per layer
      depth_multiplier: multiplier to number of repeats per stage

    úds_r1_k3_s1_e1_c16_se0.25úir_r2_k3_s2_e6_c24_se0.25úir_r2_k5_s2_e6_c40_se0.25úir_r3_k3_s2_e6_c80_se0.25úir_r3_k5_s1_e6_c112_se0.25úir_r4_k5_s2_e6_c192_se0.25úir_r1_k3_s1_e6_c320_se0.25©ró   Údivisor©r  r.   r0   Úswishr>   N©r4   r6   r8   rA   r=   r>   rQ   ©
r   r   rz   r   r    râ   r[   r]   r   rë   )rå   r÷   r  Úchannel_divisorr  ræ   rç   rø   rA   rù   rê   s              rU   Ú_gen_efficientnetrK  Ñ  sò   € ð8 
%Ð%Ø	$Ð%Ø	$Ð%Ø	$Ð%Ø	%Ð&Ø	%Ð&Ø	%Ð&ð€Hõ �>Ð6HÐRaÐbÑbÔb€LÝð Ý" 8Ð-=È*ÐUÑUÔUØ!�\ $Ñ'Ô'ØØ!Ý# F¨GÑ4Ô4Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 óN  — dgdgdgdgdgdgg}t          t          |¬¦  «        }t          dt          |||¬¦  «         |d	¦  «        d
||                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}t          | |fi |¤Ž}	|	S )z� Creates an EfficientNet-EdgeTPU model

    Ref impl: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/edgetpu
    Úer_r1_k3_s1_e4_c24_fc24_noskipÚer_r2_k3_s2_e8_c32Úer_r4_k3_s2_e8_c48Úir_r5_k5_s2_e8_c96Úir_r4_k5_s1_e8_c144Úir_r2_k5_s2_e8_c192rò   rF  r.   r0   r>   NÚrelu©r4   r6   r8   rA   r>   r=   rQ   ©
r   r   rz   r   râ   r[   r]   r   r    rë   ©
rå   r÷   r  r  ræ   rç   rø   rA   rù   rê   s
             rU   Ú_gen_efficientnet_edgerW    sé   € ð 
*Ð*Ø	ÐØ	ÐØ	ÐØ	ÐØ	Ðð	€Hõ �>Ð6HÐIÑIÔI€LÝð Ý" 8Ð-=È*ÐUÑUÔUØ!�\ $Ñ'Ô'ØØ!Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨FÑ3Ô3ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 óR  — dgdgdgdgdgdgdgg}t          t          |¬¦  «        }t          dt          |||¬	¦  «         |d
¦  «        d||                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}t          | |fi |¤Ž}	|	S )z�Creates an EfficientNet-CondConv model.

    Ref impl: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/condconv
    r=  r>  r?  r@  zir_r3_k5_s1_e6_c112_se0.25_cc4zir_r4_k5_s2_e6_c192_se0.25_cc4zir_r1_k3_s1_e6_c320_se0.25_cc4rò   )Úexperts_multiplierr.   r0   r>   NrG  rT  rQ   rU  )
rå   r÷   r  rY  ræ   rç   rø   rA   rù   rê   s
             rU   Ú_gen_efficientnet_condconvrZ  #  sð   € ð 
%Ð%Ø	$Ð%Ø	$Ð%Ø	$Ð%Ø	)Ð*Ø	)Ð*Ø	)Ð*ð€Hõ �>Ð6HÐIÑIÔI€LÝð Ý" 8Ð-=ÐRdÐeÑeÔeØ!�\ $Ñ'Ô'ØØ!Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨GÑ4Ô4ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó>  — dgdgdgdgdgdgdgg}t          dt          ||d¬	¦  «        d
ddt          t          |¬¦  «        t	          |d¦  «        |                     dd¦  «        p#t          t          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )a£  Creates an EfficientNet-Lite model.

    Ref impl: https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet/lite
    Paper: https://arxiv.org/abs/1905.11946

    EfficientNet params
    name: (channel_multiplier, depth_multiplier, resolution, dropout_rate)
      'efficientnet-lite0': (1.0, 1.0, 224, 0.2),
      'efficientnet-lite1': (1.0, 1.1, 240, 0.2),
      'efficientnet-lite2': (1.1, 1.2, 260, 0.3),
      'efficientnet-lite3': (1.2, 1.4, 280, 0.3),
      'efficientnet-lite4': (1.4, 1.8, 300, 0.3),

    Args:
      channel_multiplier: multiplier to number of channels per layer
      depth_multiplier: multiplier to number of repeats per stage
    Úds_r1_k3_s1_e1_c16rï   Úir_r2_k5_s2_e6_c40Úir_r3_k3_s2_e6_c80r.  r  rñ   T)r  r.   r0   rò   r  r>   N©r4   r6   r8   r:   rA   r=   r>   rQ   )
rz   r   r   r   r    râ   r[   r]   r   rë   ©rå   r÷   r  ræ   rç   rø   rù   rê   s           rU   Ú_gen_efficientnet_litera  C  sã   € ð& 
ÐØ	ÐØ	ÐØ	ÐØ	ÐØ	ÐØ	Ðð€Hõ ð 	Ý" 8Ð-=ÈdÐSÑSÔSØØØÝ�^Ð8JÐKÑKÔKÝ# F¨GÑ4Ô4Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gð	ð 	ð ð	ð 	€Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 óP  — dgdgdgdgdgdgg}t          t          |d¬¦  «        }t          dt          |||¬	¦  «         |d
¦  «        d||                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}t          | |fi |¤Ž}	|	S )zÙ Creates an EfficientNet-V2 base model

    Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
    Paper: `EfficientNetV2: Smaller Models and Faster Training` - https://arxiv.org/abs/2104.00298
    Úcn_r1_k3_s1_e1_c16_skipÚer_r2_k3_s2_e4_c32Úer_r2_k3_s2_e4_c48zir_r3_k3_s2_e4_c96_se0.25zir_r5_k3_s1_e6_c112_se0.25zir_r8_k3_s2_e6_c192_se0.25r2   ©ró   Úround_limitrF  r.   r0   r>   NÚsilurT  rQ   rU  rV  s
             rU   Ú_gen_efficientnetv2_baseri  l  sì   € ð 
#Ð#Ø	ÐØ	ÐØ	$Ð%Ø	%Ð&Ø	%Ð&ð€Hõ �>Ð6HÐVXÐYÑYÔY€LÝð Ý" 8Ð-=È*ÐUÑUÔUØ!�\ $Ñ'Ô'ØØ!Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨FÑ3Ô3ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ór  — dgdgdgdgdgdgg}d}|rdg|d	<   d
g|d<   d}t          t          |¬¦  «        }	t          dt          |||¬¦  «         |	|¦  «        d|	|                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}
t          | |fi |
¤Ž}|S )a[   Creates an EfficientNet-V2 Small model

    Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
    Paper: `EfficientNetV2: Smaller Models and Faster Training` - https://arxiv.org/abs/2104.00298

    NOTE: `rw` flag sets up 'small' variant to behave like my initial v2 small model,
        before ref the impl was released.
    Úcn_r2_k3_s1_e1_c24_skipÚer_r4_k3_s2_e4_c48Úer_r4_k3_s2_e4_c64zir_r6_k3_s2_e4_c128_se0.25zir_r9_k3_s1_e6_c160_se0.25zir_r15_k3_s2_e6_c256_se0.25r.   Úer_r2_k3_s1_e1_c24r   zir_r15_k3_s2_e6_c272_se0.25r•   i   rò   rF  é   r>   Nrh  rT  rQ   rU  )rå   r÷   r  r  Úrwræ   rç   rø   r6   rA   rù   rê   s               rU   Ú_gen_efficientnetv2_srq  Š  s  € ð 
#Ð#Ø	ÐØ	ÐØ	%Ð&Ø	%Ð&Ø	&Ð'ð€Hð €LØ	ð à+Ð,ˆ�‰Ø5Ð6ˆ�‰Øˆå�>Ð6HÐIÑIÔI€LÝð Ý" 8Ð-=È*ÐUÑUÔUØ!�\ ,Ñ/Ô/ØØ!Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨FÑ3Ô3ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó<  — dgdgdgdgdgdgdgg}t          dt          |||¬¦  «        d	d
t          t          |¬¦  «        |                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}t          | |fi |¤Ž}|S )zÛ Creates an EfficientNet-V2 Medium model

    Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
    Paper: `EfficientNetV2: Smaller Models and Faster Training` - https://arxiv.org/abs/2104.00298
    Úcn_r3_k3_s1_e1_c24_skipÚer_r5_k3_s2_e4_c48Úer_r5_k3_s2_e4_c80zir_r7_k3_s2_e4_c160_se0.25zir_r14_k3_s1_e6_c176_se0.25zir_r18_k3_s2_e6_c304_se0.25zir_r5_k3_s1_e6_c512_se0.25rF  r.   ro  rò   r>   Nrh  rT  rQ   ©
rz   r   r   r   râ   r[   r]   r   r    rë   ©	rå   r÷   r  r  ræ   rç   rø   rù   rê   s	            rU   Ú_gen_efficientnetv2_mrx  ²  sà   € ð 
#Ð#Ø	ÐØ	ÐØ	%Ð&Ø	&Ð'Ø	&Ð'Ø	%Ð&ð€Hõ ð Ý" 8Ð-=È*ÐUÑUÔUØØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨FÑ3Ô3ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó<  — dgdgdgdgdgdgdgg}t          dt          |||¬¦  «        d	d
t          t          |¬¦  «        |                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}t          | |fi |¤Ž}|S )zÚ Creates an EfficientNet-V2 Large model

    Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
    Paper: `EfficientNetV2: Smaller Models and Faster Training` - https://arxiv.org/abs/2104.00298
    Úcn_r4_k3_s1_e1_c32_skipÚer_r7_k3_s2_e4_c64Úer_r7_k3_s2_e4_c96zir_r10_k3_s2_e4_c192_se0.25zir_r19_k3_s1_e6_c224_se0.25zir_r25_k3_s2_e6_c384_se0.25zir_r7_k3_s1_e6_c640_se0.25rF  r.   r0   rò   r>   Nrh  rT  rQ   rv  rw  s	            rU   Ú_gen_efficientnetv2_lr}  Ò  óà   € ð 
#Ð#Ø	ÐØ	ÐØ	&Ð'Ø	&Ð'Ø	&Ð'Ø	%Ð&ð€Hõ ð Ý" 8Ð-=È*ÐUÑUÔUØØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨FÑ3Ô3ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó<  — dgdgdgdgdgdgdgg}t          dt          |||¬¦  «        d	d
t          t          |¬¦  «        |                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}t          | |fi |¤Ž}|S )zß Creates an EfficientNet-V2 Xtra-Large model

    Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
    Paper: `EfficientNetV2: Smaller Models and Faster Training` - https://arxiv.org/abs/2104.00298
    rz  Úer_r8_k3_s2_e4_c64Úer_r8_k3_s2_e4_c96zir_r16_k3_s2_e4_c192_se0.25zir_r24_k3_s1_e6_c256_se0.25zir_r32_k3_s2_e6_c512_se0.25zir_r8_k3_s1_e6_c640_se0.25rF  r.   r0   rò   r>   Nrh  rT  rQ   rv  rw  s	            rU   Ú_gen_efficientnetv2_xlr‚  ò  r~  rW   c                 ó„  — 	 |dk    rdgdgdgdgdgdgdgg}ndgd	gd
gdgdgdgdgg}t          t          ||¬¦  «        }	t          dt          |||¬¦  «         |	d¦  «        d|	t	          |d¦  «        |                     dd¦  «        p#t          t          j        fi t          |¦  «        ¤Ždœ|¤Ž}
t          | |fi |
¤Ž}|S )aŠ  Creates an EfficientNet model.

    Ref impl: https://github.com/tensorflow/tpu/blob/master/models/official/efficientnet/efficientnet_model.py
    Paper: https://arxiv.org/abs/1905.11946

    EfficientNet params
    name: (channel_multiplier, depth_multiplier, resolution, dropout_rate)
    'efficientnet-x-b0': (1.0, 1.0, 224, 0.2),
    'efficientnet-x-b1': (1.0, 1.1, 240, 0.2),
    'efficientnet-x-b2': (1.1, 1.2, 260, 0.3),
    'efficientnet-x-b3': (1.2, 1.4, 300, 0.3),
    'efficientnet-x-b4': (1.4, 1.8, 380, 0.4),
    'efficientnet-x-b5': (1.6, 2.2, 456, 0.4),
    'efficientnet-x-b6': (1.8, 2.6, 528, 0.5),
    'efficientnet-x-b7': (2.0, 3.1, 600, 0.5),
    'efficientnet-x-b8': (2.2, 3.6, 672, 0.5),
    'efficientnet-l2': (4.3, 5.3, 800, 0.5),

    Args:
      channel_multiplier: multiplier to number of channels per layer
      depth_multiplier: multiplier to number of repeats per stage

    r   zds_r1_k3_s1_e1_c16_se0.25_d1zer_r2_k3_s2_e6_c24_se0.25_nrezer_r2_k5_s2_e6_c40_se0.25_nrer@  rA  rB  rC  zer_r2_k3_s2_e4_c24_se0.25_nrezer_r2_k5_s2_e4_c40_se0.25_nrezir_r3_k3_s2_e4_c80_se0.25rD  rF  r.   r0   rh  r>   NrH  rQ   rI  )rå   r÷   r  rJ  r  Úversionræ   rç   rø   rA   rù   rê   s               rU   Ú_gen_efficientnet_xr…    s5  € ð6ð, �!‚|€|à+Ð,Ø,Ð-Ø,Ð-Ø(Ð)Ø)Ð*Ø)Ð*Ø)Ð*ð
ˆˆð ,Ð,Ø,Ð-Ø,Ð-Ø(Ð)Ø)Ð*Ø)Ð*Ø)Ð*ð
ˆõ �>Ð6HÐRaÐbÑbÔb€LÝð Ý" 8Ð-=È*ÐUÑUÔUØ!�\ $Ñ'Ô'ØØ!Ý# F¨FÑ3Ô3Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó  — dgddgddgddgdd	gd
dgg}t          dt          |¦  «        ddt          t          |¬¦  «        |                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )zªCreates a MixNet Small model.

    Ref impl: https://github.com/tensorflow/tpu/tree/master/models/official/mnasnet/mixnet
    Paper: https://arxiv.org/abs/1907.09595
    r\  zir_r1_k3_a1.1_p1.1_s2_e6_c24zir_r1_k3_a1.1_p1.1_s1_e3_c24z ir_r1_k3.5.7_s2_e6_c40_se0.5_nswú(ir_r3_k3.5_a1.1_p1.1_s1_e6_c40_se0.5_nswz&ir_r1_k3.5.7_p1.1_s2_e6_c80_se0.25_nswz$ir_r2_k3.5_p1.1_s1_e6_c80_se0.25_nswz+ir_r1_k3.5.7_a1.1_p1.1_s1_e6_c120_se0.5_nswz-ir_r2_k3.5.7.9_a1.1_p1.1_s1_e3_c120_se0.5_nswz&ir_r1_k3.5.7.9.11_s2_e6_c200_se0.5_nswz(ir_r2_k3.5.7.9_p1.1_s1_e6_c200_se0.5_nswé   r2  rò   r>   N©r4   r6   r8   rA   r>   rQ   rõ   rö   s          rU   Ú_gen_mixnet_srŠ  e  sÔ   € ð 
Ðà	'Ð)GÐHà	+Ð-WÐXà	1Ð3YÐZà	6Ð8gÐhà	1Ð3]Ð^ð€Hõ ð Ý" 8Ñ,Ô,ØØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 ó$  — dgddgddgddgdd	gd
dgg}t          dt          ||d¬¦  «        ddt          t          |¬¦  «        |                     dd¦  «        p#t          t
          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )z±Creates a MixNet Medium-Large model.

    Ref impl: https://github.com/tensorflow/tpu/tree/master/models/official/mnasnet/mixnet
    Paper: https://arxiv.org/abs/1907.09595
    Úds_r1_k3_s1_e1_c24z ir_r1_k3.5.7_a1.1_p1.1_s2_e6_c32zir_r1_k3_a1.1_p1.1_s1_e3_c32z"ir_r1_k3.5.7.9_s2_e6_c40_se0.5_nswr‡  z!ir_r1_k3.5.7_s2_e6_c80_se0.25_nswz-ir_r3_k3.5.7.9_a1.1_p1.1_s1_e6_c80_se0.25_nswzir_r1_k3_s1_e6_c120_se0.5_nswz-ir_r3_k3.5.7.9_a1.1_p1.1_s1_e3_c120_se0.5_nswz#ir_r1_k3.5.7.9_s2_e6_c200_se0.5_nswz(ir_r3_k3.5.7.9_p1.1_s1_e6_c200_se0.5_nswÚround©Údepth_truncrˆ  ro  rò   r>   Nr‰  rQ   rõ   r`  s           rU   Ú_gen_mixnet_mr�  †  sÜ   € ð 
Ðà	+Ð-KÐLà	-Ð/YÐZà	,Ð.]Ð^à	(Ð*YÐZà	.Ð0ZÐ[ð€Hõ ð Ý" 8Ð-=È7ÐSÑSÔSØØÝ�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 óz  — dgdgdgdgdgdgdgg}t          dt          ||d¬	¦  «        t          d
t          d
|dd¦  «        ¦  «        ddt	          t          |¬¦  «        t          |d¦  «        |                     dd¦  «        p#t	          t          j        fi t          |¦  «        ¤Ždœ|¤Ž}t          | |fi |¤Ž}|S )zCreates a TinyNet model.
    r=  r>  r?  r@  rA  rB  rC  r�  rŽ  r.   r
  Nr0   Trò   rG  r>   r_  rQ   )rz   r   r  r   r   r    râ   r[   r]   r   rë   )rå   Úmodel_widthr  ræ   rç   rø   rù   rê   s           rU   Ú_gen_tinynetr“  §  sú   € ð 
%Ð%Ð(CÐ'DØ	$Ð%Ð(CÐ'DØ	%Ð&Ð)EÐ(FØ	%Ð&ð	€Hõ ð 	Ý" 8Ð-=È7ÐSÑSÔSÝ˜�~¨d°KÀÀDÑIÔIÑJÔJØØÝ�^¸ÐDÑDÔDÝ# F¨GÑ4Ô4Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gð	ð 	ð ð	ð 	€Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 óD  — d| v rtd}d}d}d}t          |d¦  «        }	dt          t                   dt          fd„}
d	| v r	d
}d}g d¢}n%d| v rg d¢}nd| v rg d¢}d}nd| v rd}d}g d¢}d}nJ ‚ |
||¦  «        }n+d
}d}d}t          |d¦  «        }	dgddgddgddgddgd d!gd"gg}t          d't	          ||¦  «        |||t          t          |¬#¦  «        |                     d$d%¦  «        p#t          t          j	        fi t          |¦  «        ¤Ž|	d&œ|¤Ž}t          | |fi |¤Ž}|S )(z–
    Based on definitions in: https://github.com/tensorflow/models/tree/d2427a562f401c9af118e47af2f030a0a5599f55/official/projects/edgetpu/vision
    Ú
edgetpu_v2é@   é   r.   rS  Úchsr  c           
      ó>  — d| d         › �gd| d         › �d|› d| d         › �gd| d         › �d|› d| d         › �d| d         › �d|› d| d         › �gd| d	         › �d
| d	         › �gd| d         › �d
| d         › �gd| d         › �d
| d         › �gd| d         › �ggS )NÚcn_r1_k1_s1_cr   Úer_r1_k3_s2_e8_cr   Úer_r1_k3_s1_e4_gsÚ_crJ   Úer_r1_k3_s1_e4_cr/   Úir_r3_k3_s1_e4_cÚir_r1_k3_s1_e8_crÂ   Úir_r1_k3_s2_e8_cr—  é   rQ   )r˜  r  s     rU   Ú	_arch_defz)_gen_mobilenet_edgetpu.<locals>._arch_defÉ  s  € ð *  Q¤Ð)Ð)Ð*à, C¨¤FÐ,Ð,Ð.XÀ*Ð.XÐ.XÐPSÐTUÔPVÐ.XÐ.XÐYð 0 s¨1¤vÐ/Ð/Ø>¨
Ð>Ð>°c¸!´fÐ>Ð>Ø/ s¨1¤vÐ/Ð/Ø>¨
Ð>Ð>°c¸!´fÐ>Ð>ð	ð - C¨¤FÐ,Ð,Ð.IÀÀQÄÐ.IÐ.IÐJà, C¨¤FÐ,Ð,Ð.IÀÀQÄÐ.IÐ.IÐJà, C¨¤FÐ,Ð,Ð.IÀÀQÄÐ.IÐ.IÐJà, C¨¤FÐ,Ð,Ð-ð'ð rW   Úedgetpu_v2_xsr0   r/   )r2  r0   é0   é`   é�   é    éÀ   Úedgetpu_v2_s)ro  r¥  r–  é€   r¨  r©  é   Úedgetpu_v2_m)r0   r–  éP   r¨  r©  éð   é@  i@  Úedgetpu_v2_lé   r«  )r0   r–  r¦  r©  r¯  r¬  é€  i€  FÚcn_r1_k1_s1_c16Úer_r1_k3_s2_e8_c32Úer_r3_k3_s1_e4_c32Úer_r1_k3_s2_e8_c48Úer_r3_k3_s1_e4_c48Úir_r1_k3_s2_e8_c96Úir_r3_k3_s1_e4_c96Úir_r1_k3_s1_e8_c96_noskipÚir_r1_k5_s2_e8_c160Úir_r3_k5_s1_e4_c160Úir_r1_k3_s1_e8_c192rò   r>   N)r4   r6   r8   r9   rA   r>   r=   rQ   )r    r   r¸   rz   r   r   r   râ   r[   r]   r   rë   )rå   r÷   r  ræ   rç   r8   r9   r  r6   r=   r£  Úchannelsrø   rù   rê   s                  rU   Ú_gen_mobilenet_edgetpurÀ  ¾  sõ  € ð �wÐÐØˆ	ØÐØˆ
ØˆÝ% f¨fÑ5Ô5ˆ	ð	�4¥œ9ð 	µ#ð 	ð 	ð 	ð 	ð. ˜gÐ%Ð%ØˆIØ ÐØ6Ð6Ð6ˆHˆHØ˜wÐ&Ð&Ø7Ð7Ð7ˆHˆHØ˜wÐ&Ð&Ø7Ð7Ð7ˆHØˆLˆLØ˜wÐ&Ð&Ø ÐØˆJØ7Ð7Ð7ˆHØˆLˆLà�5à�9˜X zÑ2Ô2ˆˆð ˆ	ØÐØˆÝ% f¨fÑ5Ô5ˆ	ð Ðà!Ð#7Ð8à!Ð#7Ð8à!Ð#7Ð8à(Ð*>Ð?à"Ð$9Ð:à"Ð#ð
ˆõ" ð 	Ý" 8Ð-=Ñ>Ô>Ø!ØØ)Ý�^Ð8JÐKÑKÔKØ—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gØð	ð 	ð ð	ð 	€Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   c                 óH  — dgdgdgdgdgg}t          t          |d¬¦  «        }t          dt          ||¦  «         |d¦  «        d	||                     d
d¦  «        p#t          t
          j        fi t          |¦  «        ¤Žt          |d¦  «        dœ|¤Ž}t          | |fi |¤Ž}|S )z* Minimal test EfficientNet generator.
    rc  Úer_r1_k3_s2_e4_c24Úer_r1_k3_s2_e4_c32zir_r1_k3_s2_e4_c48_se0.25zir_r1_k3_s2_e4_c64_se0.25r2   rf  r¬  ro  r>   Nrh  rT  rQ   rU  )	rå   r÷   r  ræ   rç   rø   rA   rù   rê   s	            rU   Ú_gen_test_efficientnetrÄ    sá   € ð 
#Ð#Ø	ÐØ	ÐØ	$Ð%Ø	$Ð%ð€Hõ �>Ð6HÐVXÐYÑYÔY€LÝð Ý" 8Ð-=Ñ>Ô>Ø!�\ #Ñ&Ô&ØØ!Ø—:’:˜l¨DÑ1Ô1ÐgµW½R¼^Ð5gÐ5gÍÐ_eÑOfÔOfÐ5gÐ5gÝ# F¨FÑ3Ô3ðð ð ðð €Lõ ˜7 JÐ?Ð?°,Ð?Ð?€EØ€LrW   r1   c                 ó6   — | dddddt           t          dddd	œ|¥S )
Nr-   ©r/   éà   rÇ  ©r²  r²  g      ì?Úbicubicr_   rk   z
apache-2.0)Úurlr5   Ú
input_sizeÚ	pool_sizeÚcrop_pctÚinterpolationÚmeanÚstdÚ
first_convrk   Úlicense)r
   r   )rÊ  rç   s     rU   Ú_cfgrÓ  /  s8   € à 4°}ÐSYØ¨IÝ%Õ.BØ!°Øðð ð
 $*ðð rW   zmnasnet_050.untrainedzmnasnet_075.untrainedzmnasnet_100.rmsp_in1kzhhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mnasnet_b1-74cb7081.pthztimm/)rÊ  Ú	hf_hub_idzmnasnet_140.untrainedzsemnasnet_050.untrainedzsemnasnet_075.rmsp_in1kzkhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/semnasnet_075-18710866.pthzsemnasnet_100.rmsp_in1kzhhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mnasnet_a1-d9418771.pthzsemnasnet_140.untrainedzmnasnet_small.lamb_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mnasnet_small_lamb-aff75073.pthz#mobilenetv1_100.ra4_e3600_r224_in1k)r/   r¬  r¬  gffffffî?)rÔ  rÏ  rÐ  Útest_input_sizeÚtest_crop_pctz$mobilenetv1_100h.ra4_e3600_r224_in1kz#mobilenetv1_125.ra4_e3600_r224_in1kçÍÌÌÌÌÌì?)rÔ  rÏ  rÐ  rÍ  rÕ  rÖ  zmobilenetv2_035.untrainedzmobilenetv2_050.lamb_in1kzmhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mobilenetv2_050-3d30d450.pthrÉ  )rÊ  rÔ  rÎ  zmobilenetv2_075.untrainedzmobilenetv2_100.ra_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mobilenetv2_100_ra-b33bc2c4.pthzmobilenetv2_110d.ra_in1kzqhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mobilenetv2_110d_ra-77090ade.pthzmobilenetv2_120d.ra_in1kzqhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mobilenetv2_120d_ra-5987e2ed.pthzmobilenetv2_140.ra_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mobilenetv2_140_ra-21a4e913.pthzfbnetc_100.rmsp_in1kzhhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/fbnetc_100-c345b898.pthÚbilinearzspnasnet_100.rmsp_in1kzjhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/spnasnet_100-048bc3f4.pthzefficientnet_b0.ra_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b0_ra-3dd342df.pthz#efficientnet_b0.ra4_e3600_r224_in1kz#efficientnet_b1.ra4_e3600_r240_in1k)r/   r¯  r¯  )r
  r
  )r/   é   rÙ  )rÔ  rÏ  rÐ  rË  rÍ  rÌ  rÕ  rÖ  zefficientnet_b1.ft_in1kzmhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b1-533bc792.pth)rÊ  rÔ  rÕ  rÖ  zefficientnet_b2.ra_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b2_ra-bcdf34b7.pth)rÊ  rÔ  rË  rÌ  rÕ  rÖ  zefficientnet_b3.ra2_in1kzqhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b3_ra2-cf984f9c.pth)é	   rÚ  )r/   r°  r°  zefficientnet_b4.ra2_in1kzuhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_b4_ra2_320-7eb33cd5.pth)é
   rÛ  )r/   r³  r³  z efficientnet_b5.sw_in12k_ft_in1k)r/   éÀ  rÜ  )é   rÝ  Úsquash)rÔ  rË  rÌ  rÍ  Ú	crop_modezefficientnet_b5.sw_in12k)r/   é   rà  )é   rá  i-.  )rÔ  rË  rÌ  rÍ  r5   zefficientnet_b6.untrained)r/   é  râ  )é   rã  g¾Ÿ/Ý$î?)rÊ  rË  rÌ  rÍ  zefficientnet_b7.untrained)r/   éX  rä  )é   rå  g‘í|?5^î?zefficientnet_b8.untrained)r/   é   ræ  )é   rç  gºI+‡î?zefficientnet_l2.untrained)r/   é   rè  )é   ré  g�—nƒÀî?zefficientnet_b0_gn.untrainedzefficientnet_b0_g8_gn.untrainedz"efficientnet_b0_g16_evos.untrainedzefficientnet_b3_gn.untrained)rË  rÌ  rÕ  rÍ  zefficientnet_b3_g8_gn.untrainedzefficientnet_blur_b0.untrainedz#efficientnet_h_b5.sw_r448_e450_in1k)r/   é@  rê  )rÔ  rË  rÌ  rÍ  rß  rÕ  zefficientnet_x_b3.untrainedz#efficientnet_x_b5.sw_r448_e450_in1kzefficientnet_es.ra_in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_es_ra-f111e99c.pthzefficientnet_em.ra2_in1kzqhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_em_ra2-66250f76.pthgÓMbX9ì?)rÊ  rÔ  rË  rÌ  rÍ  zefficientnet_el.ra_in1kzmhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_el-3b455510.pth)r/   é,  rë  g!°rh‘íì?zefficientnet_es_pruned.in1kzvhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_es_pruned75-1b7248cf.pthzefficientnet_el_pruned.in1kzvhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_el_pruned70-ef2a2ccf.pthzefficientnet_cc_b0_4e.untrainedzefficientnet_cc_b0_8e.untrainedzefficientnet_cc_b1_8e.untrained)rË  rÌ  rÍ  zefficientnet_lite0.ra_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_lite0_ra-37913777.pthzefficientnet_lite1.untrainedzefficientnet_lite2.untrained)r/   é  rì  g{®Gázì?zefficientnet_lite3.untrainedzefficientnet_lite4.untrained)r/   é|  rí  )é   rî  g/Ý$�í?zefficientnet_b1_pruned.in1kzmhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tresnet/effnetb1_pruned-bea43a3a.pth)rÊ  rÔ  rË  rÌ  rÍ  rÏ  rÐ  zefficientnet_b2_pruned.in1kzmhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tresnet/effnetb2_pruned-08c1b27c.pthzefficientnet_b3_pruned.in1kzmhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-tresnet/effnetb3_pruned-59ecf72d.pthzefficientnetv2_rw_t.ra2_in1kzrhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnetv2_t_agc-3620981a.pthrÆ  rÈ  )rÊ  rÔ  rË  rÕ  rÌ  rÍ  zgc_efficientnetv2_rw_t.agc_in1kzxhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/gc_efficientnetv2_rw_t_agc-927a0bde.pthzefficientnetv2_rw_s.ra2_in1kzvhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnet_v2s_ra2_288-a6477665.pthzefficientnetv2_rw_m.agc_in1kzuhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/efficientnetv2_rw_m_agc-3d90cb1e.pthzefficientnetv2_s.untrained)rË  rÕ  rÌ  rÍ  zefficientnetv2_m.untrainedzefficientnetv2_l.untrained)r/   éà  rï  zefficientnetv2_xl.untrained)r/   é   rð  ztf_efficientnet_b0.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b0_ns-c0e6a31c.pth)rÊ  rÔ  rË  ztf_efficientnet_b1.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b1_ns-99dd0c41.pthztf_efficientnet_b2.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b2_ns-00306e48.pthztf_efficientnet_b3.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b3_ns-9d44bf68.pthztf_efficientnet_b4.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b4_ns-d6313a46.pthztf_efficientnet_b5.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b5_ns-6f26d0cf.pth)r/   éÈ  rñ  )é   rò  gÙÎ÷Sãí?ztf_efficientnet_b6.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b6_ns-51548356.pthztf_efficientnet_b7.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b7_ns-1dbc32de.pthz"tf_efficientnet_l2.ns_jft_in1k_475zwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_l2_ns_475-bebbd00a.pth)r/   éÛ  ró  gÁÊ¡E¶óí?ztf_efficientnet_l2.ns_jft_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_l2_ns-df73bb44.pthg¸…ëQ¸î?ztf_efficientnet_b0.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b0_ap-f262efe1.pth)rÊ  rÔ  rÏ  rÐ  rË  ztf_efficientnet_b1.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b1_ap-44ef0a3d.pth)rÊ  rÔ  rÏ  rÐ  rË  rÌ  rÍ  ztf_efficientnet_b2.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b2_ap-2f8e7636.pthztf_efficientnet_b3.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b3_ap-aad25bdd.pthztf_efficientnet_b4.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b4_ap-dedb23e6.pthztf_efficientnet_b5.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b5_ap-9e82fae8.pthztf_efficientnet_b6.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b6_ap-4ffb161f.pthztf_efficientnet_b7.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b7_ap-ddb28fec.pthztf_efficientnet_b8.ap_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b8_ap-00e169fa.pthztf_efficientnet_b5.ra_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b5_ra-9a3e5369.pthztf_efficientnet_b7.ra_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b7_ra-6c08e654.pthztf_efficientnet_b8.ra_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b8_ra-572d5dd9.pthztf_efficientnet_b0.aa_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b0_aa-827b6e33.pthztf_efficientnet_b1.aa_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b1_aa-ea7a6ee0.pthztf_efficientnet_b2.aa_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b2_aa-60c94f97.pthztf_efficientnet_b3.aa_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b3_aa-84b4657e.pthztf_efficientnet_b4.aa_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b4_aa-818f208c.pthztf_efficientnet_b5.aa_in1kzuhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b5_aa-99018a74.pthztf_efficientnet_b6.aa_in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b6_aa-80ba17e4.pthztf_efficientnet_b7.aa_in1kzuhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b7_aa-076e3472.pthztf_efficientnet_b0.in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b0-0af12548.pthztf_efficientnet_b1.in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b1-5c1377c4.pthztf_efficientnet_b2.in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b2-e393ef04.pthztf_efficientnet_b3.in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b3-e3bd6955.pthztf_efficientnet_b4.in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b4-74ee3bed.pthztf_efficientnet_b5.in1kzrhttps://github.com/huggingface/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_b5-c6949ce9.pthztf_efficientnet_es.in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_es-ca1afbfe.pth)ç      à?rô  rô  ztf_efficientnet_em.in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_em-e78cfe58.pthztf_efficientnet_el.in1kzphttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_el-5143854e.pthztf_efficientnet_cc_b0_4e.in1kzvhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_cc_b0_4e-4362b6b2.pth)rÊ  rÔ  rÏ  rÐ  ztf_efficientnet_cc_b0_8e.in1kzvhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_cc_b0_8e-66184a25.pthztf_efficientnet_cc_b1_8e.in1kzvhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_cc_b1_8e-f7c79ae1.pthztf_efficientnet_lite0.in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite0-0aa007d2.pth)rÊ  rÔ  rÏ  rÐ  rÎ  ztf_efficientnet_lite1.in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite1-bde8b488.pth)rÊ  rÔ  rÏ  rÐ  rË  rÌ  rÍ  rÎ  ztf_efficientnet_lite2.in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite2-dcccb7df.pthztf_efficientnet_lite3.in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite3-b733e338.pthztf_efficientnet_lite4.in1kzshttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_efficientnet_lite4-741542c3.pthgq=
×£pí?z!tf_efficientnetv2_s.in21k_ft_in1kz~https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_s_21ft1k-d7dafa41.pth)rÊ  rÔ  rÏ  rÐ  rË  rÕ  rÌ  rÍ  z!tf_efficientnetv2_m.in21k_ft_in1kz~https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_m_21ft1k-bf41664a.pth)	rÊ  rÔ  rÏ  rÐ  rË  rÕ  rÌ  rÍ  rß  z!tf_efficientnetv2_l.in21k_ft_in1kz~https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_l_21ft1k-60127a9d.pthz"tf_efficientnetv2_xl.in21k_ft_in1kz�https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_xl_in21ft1k-06c35c48.pthztf_efficientnetv2_s.in1kzwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_s-eb54923e.pthztf_efficientnetv2_m.in1kzwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_m-cc09e0cd.pthztf_efficientnetv2_l.in1kzwhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_l-d664b728.pthztf_efficientnetv2_s.in21kz{https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_s_21k-6337ad01.pthiSU  )	rÊ  rÔ  rÏ  rÐ  r5   rË  rÕ  rÌ  rÍ  ztf_efficientnetv2_m.in21kz{https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_m_21k-361418a2.pth)
rÊ  rÔ  rÏ  rÐ  r5   rË  rÕ  rÌ  rÍ  rß  ztf_efficientnetv2_l.in21kz{https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_l_21k-91a19ec9.pthztf_efficientnetv2_xl.in21kz~https://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_xl_in21k-fd7e8abf.pthztf_efficientnetv2_b0.in1kzxhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_b0-c7cc451f.pth)r/   r©  r©  )r¢  r¢  )rÊ  rÔ  rË  rÕ  rÌ  ztf_efficientnetv2_b1.in1kzxhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_b1-be6e41b0.pthztf_efficientnetv2_b2.in1kzxhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_b2-847de54e.pth)r/   éÐ   rõ  z"tf_efficientnetv2_b3.in21k_ft_in1k)rÔ  rÏ  rÐ  rË  rÕ  rÌ  rÍ  rß  ztf_efficientnetv2_b3.in1kzxhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-effv2-weights/tf_efficientnetv2_b3-57773f13.pthztf_efficientnetv2_b3.in21k)rÔ  rÏ  rÐ  r5   rË  rÕ  rÌ  rÍ  zmixnet_s.ft_in1kzfhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mixnet_s-a907afbc.pthzmixnet_m.ft_in1kzfhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mixnet_m-4647fc68.pthzmixnet_l.ft_in1kzfhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mixnet_l-5a9a2ed8.pthzmixnet_xl.ra_in1kzjhttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/mixnet_xl_ra-aac3c00c.pthzmixnet_xxl.untrainedztf_mixnet_s.in1kzihttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_mixnet_s-89d3354b.pthztf_mixnet_m.in1kzihttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_mixnet_m-0f4d8805.pthztf_mixnet_l.in1kzihttps://github.com/rwightman/pytorch-image-models/releases/download/v0.1-weights/tf_mixnet_l-6c92e0c8.pthztinynet_a.in1kzRhttps://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_a.pth)rË  rÌ  rÊ  rÔ  ztinynet_b.in1k)r/   é¼   rö  zRhttps://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_b.pthztinynet_c.in1k)r/   é¸   r÷  zRhttps://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_c.pthztinynet_d.in1k)r/   é˜   rø  )r—  r—  zRhttps://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_d.pthztinynet_e.in1k)r/   éj   rù  )rÂ   rÂ   zRhttps://github.com/huawei-noah/CV-Backbones/releases/download/v1.2.0/tinynet_e.pthzmobilenet_edgetpu_100.untrained)rË  rÍ  z!mobilenet_edgetpu_v2_xs.untrainedz mobilenet_edgetpu_v2_s.untrainedz*mobilenet_edgetpu_v2_m.ra4_e3600_r224_in1kz mobilenet_edgetpu_v2_l.untrainedztest_efficientnet.r160_in1k)r/   r¨  r¨  )rÔ  rË  rÌ  rÍ  ztest_efficientnet_ln.r160_in1kztest_efficientnet_gn.r160_in1k)rÔ  rÏ  rÐ  rË  rÌ  rÍ  z test_efficientnet_evos.r160_in1krE   c                 ó"   — t          dd| i|¤Ž}|S )z& MNASNet B1, depth multiplier of 0.5. Úmnasnet_050rô  ræ   )rû  rô  ©r  ©ræ   rç   rê   s      rU   rû  rû  …  ó!   € õ ÐPÐP¸:ÐPÈÐPÐP€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z' MNASNet B1, depth multiplier of 0.75. Úmnasnet_075ç      è?ræ   )r   r  rü  rý  s      rU   r   r   Œ  s!   € õ ÐQÐQ¸JÐQÈ&ÐQÐQ€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z& MNASNet B1, depth multiplier of 1.0. Úmnasnet_100rì   ræ   )r  rì   rü  rý  s      rU   r  r  “  rþ  rW   c                 ó"   — t          dd| i|¤Ž}|S )z& MNASNet B1,  depth multiplier of 1.4 Úmnasnet_140çffffffö?ræ   )r  r  rü  rý  s      rU   r  r  š  rþ  rW   c                 ó"   — t          dd| i|¤Ž}|S )z- MNASNet A1 (w/ SE), depth multiplier of 0.5 Úsemnasnet_050rô  ræ   )r  rô  ©rú   rý  s      rU   r  r  ¡  ó!   € õ ÐRÐR¸ZÐRÈ6ÐRÐR€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z0 MNASNet A1 (w/ SE),  depth multiplier of 0.75. Úsemnasnet_075r  ræ   )r  r  r	  rý  s      rU   r  r  ¨  s!   € õ ÐSÐS¸jÐSÈFÐSÐS€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z. MNASNet A1 (w/ SE), depth multiplier of 1.0. Úsemnasnet_100rì   ræ   )r  rì   r	  rý  s      rU   r  r  ¯  r
  rW   c                 ó"   — t          dd| i|¤Ž}|S )z. MNASNet A1 (w/ SE), depth multiplier of 1.4. Úsemnasnet_140r  ræ   )r  r  r	  rý  s      rU   r  r  ¶  r
  rW   c                 ó"   — t          dd| i|¤Ž}|S )z* MNASNet Small,  depth multiplier of 1.0. Úmnasnet_smallrì   ræ   )r  rì   )r  rý  s      rU   r  r  ½  s!   € õ ÐUÐUÀ
ÐUÈfÐUÐU€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )ú MobileNet V1 Úmobilenetv1_100rì   ræ   )r  rì   ©r  rý  s      rU   r  r  Ä  ó!   € õ ÐVÐVÀÐVÈvÐVÐV€EØ€LrW   c                 ó$   — t          dd| dœ|¤Ž}|S )r  Úmobilenetv1_100hrì   T)rÞ   ræ   )r  rì   r  rý  s      rU   r  r  Ë  s%   € õ ÐgÀÐR\ÐgÐgÐ`fÐgÐg€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )r  Úmobilenetv1_125ç      ô?ræ   )r  r  r  rý  s      rU   r  r  Ò  ó"   € õ ÐWÐWÀ*ÐWÐPVÐWÐW€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z) MobileNet V2 w/ 0.35 channel multiplier Úmobilenetv2_035çffffffÖ?ræ   )r  r   ©r"  rý  s      rU   r  r  Ù  r  rW   c                 ó"   — t          dd| i|¤Ž}|S )z( MobileNet V2 w/ 0.5 channel multiplier Úmobilenetv2_050rô  ræ   )r#  rô  r!  rý  s      rU   r#  r#  à  r  rW   c                 ó"   — t          dd| i|¤Ž}|S )z) MobileNet V2 w/ 0.75 channel multiplier Úmobilenetv2_075r  ræ   )r%  r  r!  rý  s      rU   r%  r%  ç  r  rW   c                 ó"   — t          dd| i|¤Ž}|S )z( MobileNet V2 w/ 1.0 channel multiplier Úmobilenetv2_100rì   ræ   )r'  rì   r!  rý  s      rU   r'  r'  î  r  rW   c                 ó"   — t          dd| i|¤Ž}|S )z( MobileNet V2 w/ 1.4 channel multiplier Úmobilenetv2_140r  ræ   )r)  r  r!  rý  s      rU   r)  r)  õ  r  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z3 MobileNet V2 w/ 1.1 channel, 1.2 depth multipliersÚmobilenetv2_110dçš™™™™™ñ?ç333333ó?T©r  r  ræ   )r+  r,  r!  rý  s      rU   r+  r+  ü  ó:   € õ ØðlØ25ÀTÐV`ðlð lØdjðlð l€Eà€LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z4 MobileNet V2 w/ 1.2 channel, 1.4 depth multipliers Úmobilenetv2_120dr-  r  Tr.  )r1  r-  r!  rý  s      rU   r1  r1    r/  rW   c                 ó\   — | r|                      dt          ¦  «         t          dd| i|¤Ž}|S )z	 FBNet-C Úbn_epsÚ
fbnetc_100rì   ræ   )r4  rì   )Ú
setdefaultr!   r3  rý  s      rU   r4  r4    s@   € ð ð 7à×Ò˜(Õ$5Ñ6Ô6Ð6ÝÐKÐK°jÐKÀFÐKÐK€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z Single-Path NAS Pixel1Úspnasnet_100rì   ræ   )r7  rì   )r;  rý  s      rU   r7  r7    s!   € õ ÐOÐO¸*ÐOÈÐOÐO€EØ€LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-B0 Úefficientnet_b0rì   ©r÷   r  ræ   )r9  ©rK  rý  s      rU   r9  r9    ó:   € õ ØðjØ.1ÀCÐT^ðjð jØbhðjð j€Eà€LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-B1 Úefficientnet_b1rì   r,  r:  )r>  r;  rý  s      rU   r>  r>  &  r<  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-B2 Úefficientnet_b2r,  r-  r:  )r@  r;  rý  s      rU   r@  r@  /  r<  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )ú EfficientNet-B3 Úefficientnet_b3r-  r  r:  )rC  r;  rý  s      rU   rC  rC  8  r<  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-B4 Úefficientnet_b4r  çÍÌÌÌÌÌü?r:  )rE  r;  rý  s      rU   rE  rE  A  r<  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )ú EfficientNet-B5 Úefficientnet_b5çš™™™™™ù?çš™™™™™@r:  )rI  r;  rý  s      rU   rI  rI  J  r<  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-B6 Úefficientnet_b6rF  çÍÌÌÌÌÌ@r:  )rM  r;  rý  s      rU   rM  rM  S  r<  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-B7 Úefficientnet_b7ç       @çÍÌÌÌÌÌ@r:  )rP  r;  rý  s      rU   rP  rP  \  r<  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-B8 Úefficientnet_b8rK  çÍÌÌÌÌÌ@r:  )rT  r;  rý  s      rU   rT  rT  e  r<  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-L2.Úefficientnet_l2ç333333@ç333333@r:  )rW  r;  rý  s      rU   rW  rW  n  r<  rW   c                 óN   — t          	 dt          t          d¬¦  «        | dœ|¤Ž}|S )z EfficientNet-B0 + GroupNormÚefficientnet_b0_gnr
  rF  )r>   ræ   )r[  ©rK  r   r   rý  s      rU   r[  r[  x  sE   € õ ØðoÝ)0µÈ!Ð)LÑ)LÔ)LÐYcðoð oØgmðoð o€Eà€LrW   c                 óP   — t          	 ddt          t          d¬¦  «        | dœ|¤Ž}|S )z* EfficientNet-B0 w/ group conv + GroupNormÚefficientnet_b0_g8_gnr
  rF  )r  r>   ræ   )r^  r\  rý  s      rU   r^  r^  €  sE   € õ Øð)Ø,-½'Å,Ð[\Ð:]Ñ:]Ô:]Øð)ð )à!'ð)ð )€Eð €LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z+ EfficientNet-B0 w/ group 16 conv + EvoNormÚefficientnet_b0_g16_evosr2  )r  rJ  ræ   )r`  r;  rý  s      rU   r`  r`  ‰  s7   € õ Ø"ð)Ø/1À2Øð)ð )à!'ð)ð )€Eð €LrW   c           
      óT   — t          	 ddddt          t          d¬¦  «        | dœ|¤Ž}|S )z EfficientNet-B3 w/ GroupNorm Úefficientnet_b3_gnr-  r  r2  rF  )r÷   r  rJ  r>   ræ   )rb  r\  rý  s      rU   rb  rb  ’  sN   € õ ØðZØ14ÀsÐ\^Ý�<°BÐ7Ñ7Ô7ÀJðZð ZàRXðZð Z€Eð €LrW   c                 óV   — t          	 dddddt          t          d¬¦  «        | dœ|¤Ž}|S )	z% EfficientNet-B3 w/ grouped conv + BNÚefficientnet_b3_g8_gnr-  r  r
  r2  rF  )r÷   r  r  rJ  r>   ræ   )rd  r\  rý  s      rU   rd  rd  œ  sQ   € õ ØðZØ47È#ÐZ[ÐmoÝ�<°BÐ7Ñ7Ô7ÀJðZð ZàRXðZð Z€Eð €LrW   c                 ó*   — t          	 ddd| ddœ|¤Ž}|S )z EfficientNet-B0 w/ BlurPool Úefficientnet_blur_b0rì   Úblurpc)r÷   r  ræ   r?   )rf  r;  rý  s      rU   rf  rf  ¦  s:   € õ ØðØ36ÈÐYcØðð à#ðð €Eð €LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-Edge Small. Úefficientnet_esrì   r:  )ri  ©rW  rý  s      rU   ri  ri  ±  ó:   € õ #ØðjØ.1ÀCÐT^ðjð jØbhðjð j€Eà€LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )zw EfficientNet-Edge Small Pruned. For more info: https://github.com/DeGirum/pruned-models/releases/tag/efficientnet_v1.0Úefficientnet_es_prunedrì   r:  )rm  rj  rý  s      rU   rm  rm  ¹  ó:   € õ #Ø ðqØ58È3Ð[eðqð qØioðqð q€Eà€LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-Edge-Medium. Úefficientnet_emrì   r,  r:  )rp  rj  rý  s      rU   rp  rp  À  rk  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )z EfficientNet-Edge-Large. Úefficientnet_elr-  r  r:  )rr  rj  rý  s      rU   rr  rr  È  rk  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )zw EfficientNet-Edge-Large pruned. For more info: https://github.com/DeGirum/pruned-models/releases/tag/efficientnet_v1.0Úefficientnet_el_prunedr-  r  r:  )rt  rj  rý  s      rU   rt  rt  Ï  rn  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )ú' EfficientNet-CondConv-B0 w/ 8 Experts Úefficientnet_cc_b0_4erì   r:  )rw  ©rZ  rý  s      rU   rw  rw  Ö  s:   € õ 'ØðpØ47È#ÐZdðpð pØhnðpð p€Eà€LrW   c                 ó*   — t          	 dddd| dœ|¤Ž}|S )rv  Úefficientnet_cc_b0_8erì   rJ   ©r÷   r  rY  ræ   )rz  rx  rý  s      rU   rz  rz  ß  ó:   € õ 'Øð)Ø47È#ÐbcØð)ð )à!'ð)ð )€Eð €LrW   c                 ó*   — t          	 dddd| dœ|¤Ž}|S )z' EfficientNet-CondConv-B1 w/ 8 Experts Úefficientnet_cc_b1_8erì   r,  rJ   r{  )r~  rx  rý  s      rU   r~  r~  é  r|  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )ú EfficientNet-Lite0 Úefficientnet_lite0rì   r:  )r�  ©ra  rý  s      rU   r�  r�  ó  ó:   € õ #ØðmØ14ÀsÐWaðmð mØekðmð m€Eà€LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )ú EfficientNet-Lite1 Úefficientnet_lite1rì   r,  r:  )r†  r‚  rý  s      rU   r†  r†  ü  rƒ  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )ú EfficientNet-Lite2 Úefficientnet_lite2r,  r-  r:  )r‰  r‚  rý  s      rU   r‰  r‰  	  rƒ  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )ú EfficientNet-Lite3 Úefficientnet_lite3r-  r  r:  )rŒ  r‚  rý  s      rU   rŒ  rŒ  	  rƒ  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )ú EfficientNet-Lite4 Úefficientnet_lite4r  rF  r:  )r�  r‚  rý  s      rU   r�  r�  	  rƒ  rW   c                 ó�   — |                      dt          ¦  «         |                      dd¦  «         d}t          |fddd| dœ|¤Ž}|S )	zc EfficientNet-B1 Pruned. The pruning has been obtained using https://arxiv.org/pdf/2002.08258.pdf  r3  r<   ÚsameÚefficientnet_b1_prunedrì   r,  T©r÷   r  Úprunedræ   ©r5  r!   rK  )ræ   rç   rå   rê   s       rU   r’  r’   	  sn   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ø&€GÝØðmØ$'¸#ÀdÐWaðmð mØekðmð m€Eà€LrW   c                 óŒ   — |                      dt          ¦  «         |                      dd¦  «         t          	 d	ddd| dœ|¤Ž}|S )
zb EfficientNet-B2 Pruned. The pruning has been obtained using https://arxiv.org/pdf/2002.08258.pdf r3  r<   r‘  Úefficientnet_b2_prunedr,  r-  Tr“  )r—  r•  rý  s      rU   r—  r—  +	  óg   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)ÝØ ð)Ø58È3ÐW[Øð)ð )à!'ð)ð )€Eð €LrW   c                 óŒ   — |                      dt          ¦  «         |                      dd¦  «         t          	 d	ddd| dœ|¤Ž}|S )
zb EfficientNet-B3 Pruned. The pruning has been obtained using https://arxiv.org/pdf/2002.08258.pdf r3  r<   r‘  Úefficientnet_b3_prunedr-  r  Tr“  )rš  r•  rý  s      rU   rš  rš  6	  r˜  rW   c                 ó*   — t          	 dddd| dœ|¤Ž}|S )z; EfficientNet-V2 Tiny (Custom variant, tiny not in paper). Úefficientnetv2_rw_tçš™™™™™é?r×  F©r÷   r  rp  ræ   )rœ  ©rq  rý  s      rU   rœ  rœ  A	  s=   € õ "ØðxØ25ÈÐPUÐblðxð xØpvðxð x€Eà€LrW   c           	      ó,   — t          	 ddddd| dœ|¤Ž}|S )zR EfficientNet-V2 Tiny w/ Global Context Attn (Custom variant, tiny not in paper). Úgc_efficientnetv2_rw_tr�  r×  FÚgc)r÷   r  rp  r@   ræ   )r¡  rŸ  rý  s      rU   r¡  r¡  I	  s@   € õ "Ø ðBØ58È3Ø˜4¨JðBð Bà:@ðBð B€Eð €LrW   c                 ó$   — t          dd| dœ|¤Ž}|S )zæ EfficientNet-V2 Small (RW variant).
    NOTE: This is my initial (pre official code release) w/ some differences.
    See efficientnetv2_s and tf_efficientnetv2_s for versions that match the official w/ PyTorch vs TF padding
    Úefficientnetv2_rw_sT)rp  ræ   )r¤  rŸ  rý  s      rU   r¤  r¤  R	  s$   € õ "Ðb¸DÈZÐbÐbÐ[aÐbÐb€EØ€LrW   c                 ó*   — t          	 dddd| dœ|¤Ž}|S )z* EfficientNet-V2 Medium (RW variant).
    Úefficientnetv2_rw_mr-  )r-  r-  r-  r-  rJ  rJ  Trž  )r¦  rŸ  rý  s      rU   r¦  r¦  \	  s;   € õ "Øð)Ø25ÐH_ÐdhØð)ð )à!'ð)ð )€Eð €LrW   c                 ó"   — t          dd| i|¤Ž}|S )z EfficientNet-V2 Small. Úefficientnetv2_sræ   )r¨  rŸ  rý  s      rU   r¨  r¨  f	  ó!   € õ "ÐVÐVÀÐVÈvÐVÐV€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z EfficientNet-V2 Medium. Úefficientnetv2_mræ   )r«  )rx  rý  s      rU   r«  r«  m	  r©  rW   c                 ó"   — t          dd| i|¤Ž}|S )z EfficientNet-V2 Large. Úefficientnetv2_lræ   )r­  )r}  rý  s      rU   r­  r­  t	  r©  rW   c                 ó"   — t          dd| i|¤Ž}|S )z EfficientNet-V2 Xtra-Large. Úefficientnetv2_xlræ   )r¯  )r‚  rý  s      rU   r¯  r¯  {	  s"   € õ #ÐXÐXÀ:ÐXÐQWÐXÐX€EØ€LrW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )z1 EfficientNet-B0. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnet_b0rì   r:  )r±  r•  rý  s      rU   r±  r±  ‚	  óg   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)ÝØðmØ14ÀsÐWaðmð mØekðmð m€Eà€LrW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z1 EfficientNet-B1. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnet_b1rì   r,  r:  )r´  r•  rý  s      rU   r´  r´  Œ	  r²  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z1 EfficientNet-B2. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnet_b2r,  r-  r:  )r¶  r•  rý  s      rU   r¶  r¶  –	  r²  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z0 EfficientNet-B3. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_b3r-  r  r:  )r¸  r•  rý  s      rU   r¸  r¸   	  r²  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z0 EfficientNet-B4. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_b4r  rF  r:  )rº  r•  rý  s      rU   rº  rº  ª	  r²  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z0 EfficientNet-B5. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_b5rJ  rK  r:  )r¼  r•  rý  s      rU   r¼  r¼  ´	  r²  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z0 EfficientNet-B6. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_b6rF  rN  r:  )r¾  r•  rý  s      rU   r¾  r¾  ¾	  óg   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)ÝØðmØ14ÀsÐWaðmð mØekðmð m€Eà€LrW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z0 EfficientNet-B7. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_b7rQ  rR  r:  )rÁ  r•  rý  s      rU   rÁ  rÁ  É	  r¿  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z0 EfficientNet-B8. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_b8rK  rU  r:  )rÃ  r•  rý  s      rU   rÃ  rÃ  Ô	  r¿  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z= EfficientNet-L2 NoisyStudent. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_l2rX  rY  r:  )rÅ  r•  rý  s      rU   rÅ  rÅ  ß	  r¿  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )z9 EfficientNet-Edge Small. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnet_esrì   r:  )rÇ  ©r5  r!   rW  rý  s      rU   rÇ  rÇ  ê	  óg   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ý"ØðmØ14ÀsÐWaðmð mØekðmð m€Eà€LrW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z: EfficientNet-Edge-Medium. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnet_emrì   r,  r:  )rË  rÈ  rý  s      rU   rË  rË  ô	  rÉ  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z9 EfficientNet-Edge-Large. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnet_elr-  r  r:  )rÍ  rÈ  rý  s      rU   rÍ  rÍ  þ	  rÉ  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )zF EfficientNet-CondConv-B0 w/ 4 Experts. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_cc_b0_4erì   r:  )rÏ  ©r5  r!   rZ  rý  s      rU   rÏ  rÏ  
  sg   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ý&Ø"ðsØ7:ÈSÐ]gðsð sØkqðsð s€Eà€LrW   c                 óŒ   — |                      dt          ¦  «         |                      dd¦  «         t          	 dddd| dœ|¤Ž}|S )	zF EfficientNet-CondConv-B0 w/ 8 Experts. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_cc_b0_8erì   rJ   r{  )rÒ  rÐ  rý  s      rU   rÒ  rÒ  
  óg   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ý&Ø"ð)Ø7:ÈSÐefØð)ð )à!'ð)ð )€Eð €LrW   c                 óŒ   — |                      dt          ¦  «         |                      dd¦  «         t          	 d	ddd| dœ|¤Ž}|S )
zF EfficientNet-CondConv-B1 w/ 8 Experts. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnet_cc_b1_8erì   r,  rJ   r{  )rÕ  rÐ  rý  s      rU   rÕ  rÕ  
  rÓ  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )r€  r3  r<   r‘  Útf_efficientnet_lite0rì   r:  )r×  ©r5  r!   ra  rý  s      rU   r×  r×  +
  óg   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ý"ØðpØ47È#ÐZdðpð pØhnðpð p€Eà€LrW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	r…  r3  r<   r‘  Útf_efficientnet_lite1rì   r,  r:  )rÛ  rØ  rý  s      rU   rÛ  rÛ  6
  rÙ  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	rˆ  r3  r<   r‘  Útf_efficientnet_lite2r,  r-  r:  )rÝ  rØ  rý  s      rU   rÝ  rÝ  A
  rÙ  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	r‹  r3  r<   r‘  Útf_efficientnet_lite3r-  r  r:  )rß  rØ  rý  s      rU   rß  rß  L
  rÙ  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	rŽ  r3  r<   r‘  Útf_efficientnet_lite4r  rF  r:  )rá  rØ  rý  s      rU   rá  rá  W
  rÙ  rW   c                 ó„   — |                      dt          ¦  «         |                      dd¦  «         t          dd| i|¤Ž}|S )z7 EfficientNet-V2 Small. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnetv2_sræ   )rã  )r5  r!   rq  rý  s      rU   rã  rã  b
  óO   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ý!ÐYÐYÀJÐYÐRXÐYÐY€EØ€LrW   c                 ó„   — |                      dt          ¦  «         |                      dd¦  «         t          dd| i|¤Ž}|S )z8 EfficientNet-V2 Medium. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnetv2_mræ   )ræ  )r5  r!   rx  rý  s      rU   ræ  ræ  k
  rä  rW   c                 ó„   — |                      dt          ¦  «         |                      dd¦  «         t          dd| i|¤Ž}|S )z7 EfficientNet-V2 Large. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnetv2_lræ   )rè  )r5  r!   r}  rý  s      rU   rè  rè  t
  rä  rW   c                 ó„   — |                      dt          ¦  «         |                      dd¦  «         t          dd| i|¤Ž}|S )z? EfficientNet-V2 Xtra-Large. Tensorflow compatible variant
    r3  r<   r‘  Útf_efficientnetv2_xlræ   )rê  )r5  r!   r‚  rý  s      rU   rê  rê  }
  sO   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ý"Ð[Ð[ÀjÐ[ÐTZÐ[Ð[€EØ€LrW   c                 ó„   — |                      dt          ¦  «         |                      dd¦  «         t          dd| i|¤Ž}|S )z4 EfficientNet-V2-B0. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnetv2_b0ræ   )rì  ©r5  r!   ri  rý  s      rU   rì  rì  ‡
  sO   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ý$Ð]Ð]È
Ð]ÐV\Ð]Ð]€EØ€LrW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z4 EfficientNet-V2-B1. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnetv2_b1rì   r,  r:  )rï  rí  rý  s      rU   rï  rï  �
  óg   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)Ý$ØðoØ36ÈÐYcðoð oØgmðoð o€Eà€LrW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z4 EfficientNet-V2-B2. Tensorflow compatible variant  r3  r<   r‘  Útf_efficientnetv2_b2r,  r-  r:  )rò  rí  rý  s      rU   rò  rò  š
  rð  rW   c                 óŠ   — |                      dt          ¦  «         |                      dd¦  «         t          	 ddd| dœ|¤Ž}|S )	z3 EfficientNet-V2-B3. Tensorflow compatible variant r3  r<   r‘  Útf_efficientnetv2_b3r-  r  r:  )rô  rí  rý  s      rU   rô  rô  ¤
  rð  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )rB  Úefficientnet_x_b3r-  r  r:  )rö  ©r…  rý  s      rU   rö  rö  ®
  s:   € õ  ØðlØ03ÀcÐV`ðlð lØdjðlð l€Eà€LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )rH  Úefficientnet_x_b5rJ  rK  r:  )rù  r÷  rý  s      rU   rù  rù  ·
  s:   € õ  ØðlØ03ÀcÐV`ðlð lØdjðlð l€Eà€LrW   c                 ó*   — t          	 dddd| dœ|¤Ž}|S )rH  Úefficientnet_h_b5g¸…ëQ¸þ?rK  rJ   )r÷   r  r„  ræ   )rû  r÷  rý  s      rU   rû  rû  ¿
  s=   € õ  ØðxØ04ÀsÐTUÐblðxð xØpvðxð x€Eà€LrW   c                 ó&   — t          	 dd| dœ|¤Ž}|S )z"Creates a MixNet Small model.
    Úmixnet_srì   ©r÷   ræ   )rý  )rŠ  rý  s      rU   rý  rý  Ç
  ó7   € õ ØðMØ'*°zðMð MØEKðMð M€Eà€LrW   c                 ó&   — t          	 dd| dœ|¤Ž}|S )z#Creates a MixNet Medium model.
    Úmixnet_mrì   rþ  )r  ©r�  rý  s      rU   r  r  Ð
  rÿ  rW   c                 ó&   — t          	 dd| dœ|¤Ž}|S )z"Creates a MixNet Large model.
    Úmixnet_lçÍÌÌÌÌÌô?rþ  )r  r  rý  s      rU   r  r  Ù
  rÿ  rW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )zgCreates a MixNet Extra-Large model.
    Not a paper spec, experimental def by RW w/ depth scaling.
    Ú	mixnet_xlrJ  r-  r:  )r  r  rý  s      rU   r  r  â
  s9   € õ
 ØðdØ(+¸cÈjðdð dØ\bðdð d€Eà€LrW   c                 ó(   — t          	 ddd| dœ|¤Ž}|S )znCreates a MixNet Double Extra Large model.
    Not a paper spec, experimental def by RW w/ depth scaling.
    Ú
mixnet_xxlg333333@r  r:  )r	  r  rý  s      rU   r	  r	  ì
  s9   € õ
 ØðeØ),¸sÈzðeð eØ]cðeð e€Eà€LrW   c                 óˆ   — |                      dt          ¦  «         |                      dd¦  «         t          	 dd| dœ|¤Ž}|S )z@Creates a MixNet Small model. Tensorflow compatible variant
    r3  r<   r‘  Útf_mixnet_srì   rþ  )r  )r5  r!   rŠ  rý  s      rU   r  r  ö
  ód   € ð ×Ò�hÕ 1Ñ2Ô2Ð2Ø
×Ò�j &Ñ)Ô)Ð)ÝØðPØ*-¸*ðPð PØHNðPð P€Eà€LrW   c                 óˆ   — |                      dt          ¦  «         |                      dd¦  «         t          	 dd| dœ|¤Ž}|S )zACreates a MixNet Medium model. Tensorflow compatible variant
    r3  r<   r‘  Útf_mixnet_mrì   rþ  )r  ©r5  r!   r�  rý  s      rU   r  r    r  rW   c                 óˆ   — |                      dt          ¦  «         |                      dd¦  «         t          	 dd| dœ|¤Ž}|S )z@Creates a MixNet Large model. Tensorflow compatible variant
    r3  r<   r‘  Útf_mixnet_lr  rþ  )r  r  rý  s      rU   r  r    r  rW   c                 ó"   — t          dd| i|¤Ž}|S )N)Ú	tinynet_arì   r-  ræ   ©r“  rý  s      rU   r  r    s   € åÐPÐP¸:ÐPÈÐPÐP€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )N)Ú	tinynet_br  r,  ræ   r  rý  s      rU   r  r    ó   € åÐQÐQ¸JÐQÈ&ÐQÐQ€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )N)Ú	tinynet_cçHáz®Gá?g333333ë?ræ   r  rý  s      rU   r  r  #  s   € åÐRÐR¸ZÐRÈ6ÐRÐR€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )N)Ú	tinynet_dr  g=
×£p=æ?ræ   r  rý  s      rU   r  r  )  s   € åÐSÐS¸jÐSÈFÐSÐS€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )N)Ú	tinynet_egR¸…ëQà?g333333ã?ræ   r  rý  s      rU   r  r  /  r  rW   c                 ó"   — t          dd| i|¤Ž}|S )z MobileNet-EdgeTPU-v1 100. Úmobilenet_edgetpu_100ræ   )r   ©rÀ  rý  s      rU   r   r   5  s"   € õ #Ð\Ð\ÀzÐ\ÐU[Ð\Ð\€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z# MobileNet-EdgeTPU-v2 Extra Small. Úmobilenet_edgetpu_v2_xsræ   )r#  r!  rý  s      rU   r#  r#  <  s"   € õ #Ð^Ð^ÈÐ^ÐW]Ð^Ð^€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z MobileNet-EdgeTPU-v2 Small. Úmobilenet_edgetpu_v2_sræ   )r%  r!  rý  s      rU   r%  r%  C  ó"   € õ #Ð]Ð]È
Ð]ÐV\Ð]Ð]€EØ€LrW   c                 ó"   — t          dd| i|¤Ž}|S )z MobileNet-EdgeTPU-v2 Medium. Úmobilenet_edgetpu_v2_mræ   )r(  r!  rý  s      rU   r(  r(  J  r&  rW   c                 ó"   — t          dd| i|¤Ž}|S )z MobileNet-EdgeTPU-v2 Large. Úmobilenet_edgetpu_v2_lræ   )r*  r!  rý  s      rU   r*  r*  Q  r&  rW   c                 ó"   — t          dd| i|¤Ž}|S )NÚtest_efficientnetræ   )r,  )rÄ  rý  s      rU   r,  r,  X  s    € å"ÐXÐXÀ:ÐXÐQWÐXÐX€EØ€LrW   c                 óv   — t          	 d| |                     dt          t          d¬¦  «        ¦  «        dœ|¤Ž}|S )NÚtest_efficientnet_gnr>   r
  rF  ©ræ   r>   )r.  )rÄ  râ   r   r   rý  s      rU   r.  r.  ^  sR   € õ #ØðàØ—:’:˜l­GµLÈQÐ,OÑ,OÔ,OÑPÔPðð ð ð	ð €Eð €LrW   c                 óX   — t          	 d| |                     dt          ¦  «        dœ|¤Ž}|S )NÚtest_efficientnet_lnr>   r/  )r1  )rÄ  râ   r   rý  s      rU   r1  r1  j  sC   € å"ØðàØ—:’:˜l­NÑ;Ô;ðð ð ð	ð €Eð €LrW   c                 óv   — t          	 d| |                     dt          t          d¬¦  «        ¦  «        dœ|¤Ž}|S )NÚtest_efficientnet_evosr>   r
  rF  r/  )r3  )rÄ  râ   r   r   rý  s      rU   r3  r3  u  sP   € å"Ø ðàØ—:’:˜l­GµKÈAÐ,NÑ,NÔ,NÑOÔOðð ð ð	ð €Eð €LrW   Útf_efficientnet_b0_apÚtf_efficientnet_b1_apÚtf_efficientnet_b2_apÚtf_efficientnet_b3_apÚtf_efficientnet_b4_apÚtf_efficientnet_b5_apÚtf_efficientnet_b6_apÚtf_efficientnet_b7_apÚtf_efficientnet_b8_apÚtf_efficientnet_b0_nsÚtf_efficientnet_b1_nsÚtf_efficientnet_b2_nsÚtf_efficientnet_b3_nsÚtf_efficientnet_b4_nsÚtf_efficientnet_b5_nsÚtf_efficientnet_b6_nsÚtf_efficientnet_b7_nsr@  rC  r  r  )Útf_efficientnet_l2_ns_475Útf_efficientnet_l2_nsÚtf_efficientnetv2_s_in21ft1kÚtf_efficientnetv2_m_in21ft1kÚtf_efficientnetv2_l_in21ft1kÚtf_efficientnetv2_xl_in21ft1kÚtf_efficientnetv2_s_in21kÚtf_efficientnetv2_m_in21kÚtf_efficientnetv2_l_in21kÚtf_efficientnetv2_xl_in21kÚefficientnet_b2aÚefficientnet_b3aÚ
mnasnet_a1Ú
mnasnet_b1r²   )rì   F)rì   rì   NFFF)rì   rì   NFF)rì   rì   r
  NF)rì   rì   NF)rì   rì   r   F)rì   rì   F)rì   rì   r
  Nr   F)r1   )Èr·   Ú	functoolsr   Útypingr   r   r   r   r   r	   r˜   Útorch.nnr[   Útorch.nn.functionalÚ
functionalr­   Ú	timm.datar
   r   r   r   Útimm.layersr   r   r   r   r   r   r   r   Ú_builderr   r   Ú_efficientnet_blocksr   Ú_efficientnet_builderr   r   r   r   r   r   r    r!   Ú	_featuresr"   r#   r$   Ú_manipulater%   r&   Ú	_registryr'   r(   r)   Ú__all__r½   r*   r+   rë   rú   r  r  r  r"  r3  r;  rK  rW  rZ  ra  ri  rq  rx  r}  r‚  r…  rŠ  r�  r“  rÀ  rÄ  rÓ  Údefault_cfgsrû  r   r  r  r  r  r  r  r  r  r  r  r  r#  r%  r'  r)  r+  r1  r4  r7  r9  r>  r@  rC  rE  rI  rM  rP  rT  rW  r[  r^  r`  rb  rd  rf  ri  rm  rp  rr  rt  rw  rz  r~  r�  r†  r‰  rŒ  r�  r’  r—  rš  rœ  r¡  r¤  r¦  r¨  r«  r­  r¯  r±  r´  r¶  r¸  rº  r¼  r¾  rÁ  rÃ  rÅ  rÇ  rË  rÍ  rÏ  rÒ  rÕ  r×  rÛ  rÝ  rß  rá  rã  ræ  rè  rê  rì  rï  rò  rô  rö  rù  rû  rý  r  r  r  r	  r  r  r  r  r  r  r  r  r   r#  r%  r(  r*  r,  r.  r1  r3  r´   rQ   rW   rU   ú<module>rb     sÒ5  ðð$ð $ðJ Ð Ð Ð Ð Ð Ø ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?Ð ?à €€€Ø Ð Ð Ð Ð Ð Ø Ð Ð Ð Ð Ð Ð Ð Ð à rÐ rÐ rÐ rÐ rÐ rÐ rÐ rÐ rÐ rÐ rÐ rð@ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @ð @à GÐ GÐ GÐ GÐ GÐ GÐ GÐ GØ /Ð /Ð /Ð /Ð /Ð /ðJð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jð Jà FÐ FÐ FÐ FÐ FÐ FÐ FÐ FÐ FÐ FØ 3Ð 3Ð 3Ð 3Ð 3Ð 3Ð 3Ð 3Ø YÐ YÐ YÐ YÐ YÐ YÐ YÐ YÐ YÐ YàÐ1Ð
2€ðfð fð fð fð f�2”9ñ fô fð fðR	`&ð `&ð `&ð `&ð `&˜2œ9ñ `&ô `&ð `&ñFð ð ð ñ6!ð !ð !ð !ñH!ð !ð !ð !ñHð ð ð ð< ;>ØJOñ!ð !ð !ð !ðJ ;>Ø9>ñ"ð "ð "ð "ñJð ð ð ñ< ð  ð  ð  ðH PQØ$)ñ/ð /ð /ð /ðf \añð ð ð ðB afñð ð ð ñ@&ð &ð &ð &ðT \añð ð ð ð> fkñ%ð %ð %ð %ðR \añð ð ð ðB \añð ð ð ðB \añð ð ð ðB PQØ/4ñPð Pð Pð Pñfð ð ð ñBð ð ð ñBð ð ð ñ.Vð Vð Vð Vñrð ð ð ñ0ð ð ð ð %Ð$ð I	&Ø˜T˜T™VœVðI	&à˜T˜T™VœVðI	&ð ˜T˜TØvØðñ ô ðI	&ð ˜T˜T™VœVðI	&ð ˜t˜t™vœvðI	&ð ˜t˜tØyØð ñ  ô  ðI	&ð ˜t˜tØvØð ñ  ô  ðI	&ð ˜t˜t™vœvðI	&ð  ˜t˜tØ~Øð ñ  ô  ð!I	&ð( *¨4¨4ØØ$Ð*@Ø%°Tð,ñ ,ô ,ð)I	&ð2 +¨D¨DØØ$Ð*@Ø%°Tð-ñ -ô -ð3I	&ð< *¨4¨4ØØ$Ð*@Ø mÀ3ð,ñ ,ô ,ð=I	&ðH    ¡¤ðII	&ðJ    Ø{ØØð"ñ "ô "ðKI	&ðT    ¡¤ðUI	&ðV ˜t˜tØ~Øð ñ  ô  ðWI	&ð\   ØØð!ñ !ô !ð]I	&ð I	&ðb   ØØð!ñ !ô !ðcI	&ðh ˜t˜tØ~Øð ñ  ô  ðiI	&ðp ˜D˜DØvØØ ð"ñ "ô "ðqI	&ðx ˜d˜dØxØØ ð"ñ "ô "ðyI	&ðD ˜t˜tØ~Øð ñ  ô  ðEI	&ðJ *¨4¨4ØØ$Ð*@Ø mÀ3ð,Hñ ,Hô ,HðKI	&ðR *¨4¨4ØØ$Ð*@Ø ¨3¸&Ø%°Sð	,:ñ ,:ô ,:ðSI	&ð\ ˜t˜tØ{ØØ%°Sð :ñ  :ô  :ð]I	&ðd ˜t˜tØ~ØØ ¨FÀMÐadð fñ  fô  fðeI	&ðl   ØØØ ¨FÀMÐadð!fñ !fô !fðmI	&ðt   ð DØØ ¨HÀmÐcfð!hñ !hô !hðuI	&ð| '¨¨ØØ ¨H¸sÈhð)Xñ )Xô )Xð}I	&ðB   ØØ ¨H¸tÐQVð!Xñ !Xô !XðCI	&ðH    Ø˜=°HÀuð"Nñ "Nô "NðII	&ðL    Ø˜=°HÀuð"Nñ "Nô "NðMI	&ðP    Ø˜=°HÀuð"Nñ "Nô "NðQI	&ðT    Ø˜=°HÀuð"Nñ "Nô "NðUI	&ð I	&ð I	&ð\ # D D¡F¤Fð]I	&ð^ & t t¡v¤vð_I	&ð` )¨$¨$©&¬&ðaI	&ðb # D DØ ¨FÀMÐ\_ð%añ %aô %aðcI	&ðf & t tØ ¨FÀMÐ\_ð(añ (aô (aðgI	&ðj % d d¡f¤fðkI	&ðl *¨4¨4ØØ ¨H¸sØ¨Mð,;ñ ,;ô ,;ðmI	&ðt " 4 4Ø˜=°FÀTð$Kñ $Kô $KðuI	&ðx *¨4¨4ØØ ¨H¸sØ¨Mð,;ñ ,;ô ,;ðyI	&ðB ˜t˜tØ~Øð ñ  ô  ðCI	&ðH   ØØØ ¨F¸Uð!Dñ !Dô !DðII	&ðP ˜t˜tØ{ØØ ¨H¸uð Fñ  Fô  FðQI	&ðZ " 4 4ð EØð$ñ $ô $ð[I	&ð` " 4 4ð EØØ ¨H¸uð$Fñ $Fô $FðaI	&ðj & t t¡v¤vðkI	&ðl & t t¡v¤vðmI	&ðn & t t°}ÐPVÐafÐ'gÑ'gÔ'gðoI	&ð I	&ð I	&ðr ! $ $ð BØð#ñ #ô #ðsI	&ðx # D DØ ¨F¸Uð%Dñ %Dô %DðyI	&ð| # D DØ ¨F¸Uð%Dñ %Dô %Dð}I	&ð@ # D DØ ¨H¸uð%Fñ %Fô %FðAI	&ðD # D DØ ¨H¸uð%Fñ %Fô %FðEI	&ðJ " 4 4Ø{ØØ ¨FØÐ4Ð:Pð	$Rñ $Rô $RðKI	&ðT " 4 4Ø{ØØ ¨FØÐ4Ð:Pð	$Rñ $Rô $RðUI	&ð^ " 4 4Ø{ØØ ¨HØÐ4Ð:Pð	$Rñ $Rô $Rð_I	&ðj # D Dð AØØ °-È6Ð\_ð%añ %aô %aðkI	&ðr & t tð GØØ °-È6Ð\_ð(añ (aô (aðsI	&ðz # D Dð EØØ °-È6Ð\_ð%añ %aô %að{I	&ðB # D Dð DØØ °-È8Ð^að%cñ %cô %cðCI	&ðL ! $ $Ø °-È6Ð\_ð#añ #aô #aðMI	&ðP ! $ $Ø °-È8Ð^að#cñ #cô #cðQI	&ðT ! $ $Ø °-È8Ð^að#cñ #cô #cðUI	&ðX " 4 4Ø °-È8Ð^að$cñ $cô $cðYI	&ð^ % d dð BØØ ð'"ñ '"ô '"ð_I	&ð I	&ð I	&ðf % d dð BØØ ¨F¸Uð'Dñ 'Dô 'DðgI	&ðn % d dð BØØ ¨F¸Uð'Dñ 'Dô 'DðoI	&ðv % d dð BØØ ¨H¸uð'Fñ 'Fô 'FðwI	&ð~ % d dð BØØ ¨H¸uð'Fñ 'Fô 'FðI	&ðF % d dð BØØ ¨H¸uð'Fñ 'Fô 'FðGI	&ðN % d dð BØØ ¨H¸uð'Fñ 'Fô 'FðOI	&ðV % d dð BØØ ¨H¸uð'Fñ 'Fô 'FðWI	&ð^ )¨$¨$ð FØØ ¨H¸uð+Fñ +Fô +Fð_I	&ðf % d dð BØØ ¨H¸tð'Eñ 'Eô 'EðgI	&ðp ! $ $ð BØØ$Ð*@È]ð#\ñ #\ô #\ðqI	&ðx ! $ $ð BØØ$Ð*@Ø ¨F¸Uð	#Dñ #Dô #DðyI	&ðB ! $ $ð BØØ$Ð*@Ø ¨F¸Uð	#Dñ #Dô #DðCI	&ðL ! $ $ð BØØ$Ð*@Ø ¨H¸uð	#Fñ #Fô #FðMI	&ðV ! $ $ð BØØ$Ð*@Ø ¨H¸uð	#Fñ #Fô #FðWI	&ð` ! $ $ð BØØ$Ð*@Ø ¨H¸uð	#Fñ #Fô #FðaI	&ðj ! $ $ð BØØ$Ð*@Ø ¨H¸uð	#Fñ #Fô #FðkI	&ðt ! $ $ð BØØ$Ð*@Ø ¨H¸uð	#Fñ #Fô #FðuI	&ð I	&ð I	&ð~ ! $ $ð BØØ$Ð*@Ø ¨H¸uð	#Fñ #Fô #FðI	&ðJ	 ! $ $ð BØØ ¨H¸uð#Fñ #Fô #FðK	I	&ðR	 ! $ $ð BØØ ¨H¸uð#Fñ #Fô #FðS	I	&ðZ	 ! $ $ð BØØ ¨H¸uð#Fñ #Fô #Fð[	I	&ñd	 ! $ $ñ BØØ ð#"ñ #"ô #"ðe	I	&ñl	 ! $ $ñ BØØ ¨F¸Uð#Dñ #Dô #Dðm	I	&ñt	 ! $ $ñ BØØ ¨F¸Uð#Dñ #Dô #Dðu	I	&ñ|	 ! $ $ñ BØØ ¨H¸uð#Fñ #Fô #Fð}	I	&ñD
 ! $ $ñ BØØ ¨H¸uð#Fñ #Fô #FðE
I	&ñL
 ! $ $ñ DØØ ¨H¸uð#Fñ #Fô #FðM
I	&ñT
 ! $ $ñ BØØ ¨H¸uð#Fñ #Fô #FðU
I	&ñ\
 ! $ $ñ DØØ ¨H¸uð#Fñ #Fô #Fð]
I	&ñf
 ˜t˜tñ AØØ ð "ñ  "ô  "ðg
I	&ñn
 ˜t˜tñ AØØ ¨F¸Uð Dñ  Dô  Dðo
I	&ñv
 ˜t˜tñ AØØ ¨F¸Uð Dñ  Dô  Dðw
I	&ñ~
 ˜t˜tñ AØØ ¨H¸uð Fñ  Fô  Fð
I	&ñF ˜t˜tñ AØØ ¨H¸uð Fñ  Fô  FðGI	&ð I	&ð I	&ñN ˜t˜tñ AØØ ¨H¸uð Fñ  Fô  FðOI	&ñX ˜t˜tÙ~ØÙ¡/Ø ð	 $ñ  $ô  $ðYI	&ñb ˜t˜tÙ~ØÙ¡/Ø ¨F¸Uð	 Dñ  Dô  DðcI	&ñl ˜t˜tÙ~ØÙ¡/Ø ¨H¸uð	 Fñ  Fô  FðmI	&ñx $ T Tñ EØØ$Ð*@ñ&Bñ &Bô &BðyI	&ñ@ $ T Tñ EØØ$Ð*@ñ&Bñ &Bô &BðAI	&ñH $ T Tñ EØØ$Ð*@Ø ¨F¸Uð	&Dñ &Dô &DðII	&ñT ! $ $ñ BØÙ¡/Øñ	#ñ #ô #ðUI	&ñ` ! $ $ñ BØÙ¡/Ø ¨F¸UØñ#ñ #ô #ðaI	&ñn ! $ $ñ BØÙ¡/Ø ¨F¸UØñ#ñ #ô #ðoI	&ñ| ! $ $ñ BØÙ¡/Ø ¨H¸uÐT^ñ	#`ñ #`ô #`ð}I	&ñF ! $ $ñ BØÙ¡/Ø ¨H¹uÐT^ñ	#`ñ #`ô #`ðGI	&ñR (¨¨ñ MØÙ¡/Ø °-È8Ð^añ	*cñ *cô *cðSI	&ñ\ (¨¨ñ MØÙ¡/Ø °-È8Ð^aÐmuñ	*wñ *wô *wð]I	&ñf (¨¨ñ MØÙ¡/Ø °-È8Ð^aÐmuñ	*wñ *wô *wðgI	&ñp )¨$¨$ñ PØÙ¡/Ø °-È8Ð^aÐmuñ	+wñ +wô +wðqI	&ñ|   ñ FØÙ¡/Ø °-È8Ð^añ	!cñ !cô !cð}I	&ð I	&ð I	&ñF   ñ FØÙ¡/Ø °-È8Ð^aÐmuñ	!wñ !wô !wðGI	&ñP   ñ FØÙ¡/Ø °-È8Ð^aÐmuñ	!wñ !wô !wðQI	&ñ\    ñ JØÙ¡/¹uØ °-È8Ð^añ	"cñ "cô "cð]I	&ñf    ñ JØÙ¡/¹uØ °-È8Ð^aÐmuñ	"wñ "wô "wðgI	&ñp    ñ JØÙ¡/¹uØ °-È8Ð^aÐmuñ	"wñ "wô "wðqI	&ñz ! $ $ñ MØÙ¡/¹uØ °-È8Ð^aÐmuñ	#wñ #wô #wð{I	&ñF    ñ GØÙ °-É6ñ"Sñ "Sô "SðGI	&ñN    ñ GØÙ °-É6Ð\að"cñ "cô "cðOI	&ñV    ñ GØÙ °-È6Ð\að"cñ "cô "cðWI	&ñ^ )¨$¨$ØØ$Ð*@Ø °-È6Ð\_Ðksñ+uñ +uô +uð_I	&ñf    ñ GØØ °-È6Ð\að"cñ "cô "cðgI	&ñn ! $ $ØØ$Ð*@ÉeØ °-È6Ð\añ#cñ #cô #cðoI	&ñx ˜˜ÙtØðñ ô ðyI	&ñ~ ˜˜ÙtØðñ ô ðI	&ñD ˜˜ÙtØðñ ô ðEI	&ñJ ˜˜ÙxØðñ ô ðKI	&ñP ˜D˜D™FœFðQI	&ð I	&ð I	&ñT ˜˜ÙwØðñ ô ðUI	&ñZ ˜˜ÙwØðñ ô ð[I	&ñ` ˜˜ÙwØðñ ô ðaI	&ñh �d�dÙ ©FÙ`Øññ ô ðiI	&ñp �d�dÙ ©FÙ`Øññ ô ðqI	&ñx �d�dÙ ©FÙ`Øññ ô ðyI	&ñ@ �d�dÙ ©FÙ`Øññ ô ðAI	&ñH �d�dÙ ©FÙ`Øññ ô ðII	&ñR & t tà ¨3ñ(0ñ (0ô (0ðSI	&ñX (¨¨à ¨3ñ*0ñ *0ô *0ðYI	&ñ^ '¨¨à ¨3ñ)0ñ )0ô )0ð_I	&ñd 1°$°$ØØ$Ð*@Ø mÀ4ð3ñ 3ô 3ðeI	&ñn '¨¨à ¨3ñ)0ñ )0ô )0ðoI	&ñv " 4 4ØÙ ©F¸Tñ$Cñ $Cô $CðwI	&ñ| % d dØÙ ©F¸Tñ'Cñ 'Cô 'Cð}I	&ñB % d dØÙ¡/Ù ©F¸Tñ'Cñ 'Cô 'CðCI	&ñJ '¨¨ØÙ¡/Ù ©F¸Tñ)Cñ )Cô )CðKI	&ð I	&ñ I	ô I	€ðX ññ ¨|ð ñ ð ñ „ðð ññ ¨|ð ñ ð ñ „ðð ññ ¨|ð ñ ð ñ „ðð ññ ¨|ð ñ ð ñ „ðð ññ °ð ñ ð ñ „ðð ññ °ð ñ ð ñ „ðð ññ °ð ñ ð ñ „ðð ññ °ð ñ ð ñ „ðð ññ °ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °Lð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °Lð ñ ð ñ „ðð ññ °Lð ñ ð ñ „ðð ññ ¨lð ñ ð ñ „ðð ññ °ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸Lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ °<ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ °|ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ °|ð ñ ð ñ „ðð ññ °|ð ñ ð ñ „ðð ññ °Lð ñ ð ñ „ðð ññ °Lð ñ ð ñ „ðð ññ °Lð ñ ð ñ „ðð ññ °\ð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ °lð ñ ð ñ „ðð ññ ¸Lð ñ ð ñ „ðð ññ ¸Lð ñ ð ñ „ðð ññ ¸Lð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ °|ð ñ ð ñ „ðð ññ °|ð ñ ð ñ „ðð ññ °|ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ °\ð ñ ð ñ „ðð ññ °\ð ñ ð ñ „ðð ññ °\ð ñ ð ñ „ðð ññ ¨Lð ñ ð ñ „ðð ññ ¨Lð ñ ð ñ „ðð ññ ¨Lð ñ ð ñ „ðð ññ ¨\ð ñ ð ñ „ðð ññ ¨lð ñ ð ñ „ðð ññ ¨|ð ñ ð ñ „ðð ññ ¨|ð ñ ð ñ „ðð ññ ¨|ð ñ ð ñ „ðð ññ ¨\ð ñ ð ñ „ðð
 ññ ¨\ð ñ ð ñ „ðð
 ññ ¨\ð ñ ð ñ „ðð
 ññ ¨\ð ñ ð ñ „ðð
 ññ ¨\ð ñ ð ñ „ðð
 ññ ¸ð ñ ð ñ „ðð ññ ¸<ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð ññ °\ð ñ ð ñ „ðð
 ññ ¸ð ñ ð ñ „ðð ññ ¸ð ñ ð ñ „ðð ññ ¸,ð ñ ð ñ „ðð Ð ˜Hð  'ÙÐ9ð 'áÐ9ð 'ñ Ð9ð 'ñ Ð9ð	 'ñ
 Ð9ð 'ñ Ð9ð 'ñ Ð9ð 'ñ Ð9ð 'ñ Ð9ð 'ñ Ð=ð 'ñ Ð=ð 'ñ Ð=ð 'ñ Ð=ð 'ñ Ð=ð 'ñ Ð=ð 'ñ  Ð=ð! 'ñ" Ð=ð# 'ð$ "FØ=Ù$GÙ$GÙ$GÙ%IÙ!<Ù!<Ù!<Ù">Ù)Ù)Ù!Ùñ? 'ð  'ð  'ñ  ô  ð  ð  ð  rW   