jd UddlZddlmZddlmZddlmZmZmZm Z m Z m Z ddl Z ddl m Z mZddlmZmZddlmZmZdd lmZd d lmZmZmZd d lmZd d lmZmZm Z gdZ!GddeZ"Gdde j#Z$Gdde j%Z&Gdde j#Z'GddZ(Gdde j%Z)de(de ede*dede)f d Z+d!ed"Z,ee-efe.d#<ie,d$d%d&Z/Gd'd(eZ0Gd)d*eZ1Gd+d,eZ2Gd-d.eZ3Gd/d0eZ4Gd1d2eZ5Gd3d4eZ6Gd5d6eZ7Gd7d8eZ8Gd9d:eZ9Gd;deZ;Gd?d@eZ<GdAdBeZ=GdCdDeZ>ee dEe0j?fFddGdHde e0de*dede)fdIZ@ee dEe1j?fFddGdHde e1de*dede)fdJZAee dEe2j?fFddGdHde e2de*dede)fdKZBee dEe3j?fFddGdHde e3de*dede)fdLZCee dEe4j?fFddGdHde e4de*dede)fdMZDee dEe5j?fFddGdHde e5de*dede)fdNZEee dEe6j?fFddGdHde e6de*dede)fdOZFee dPFddGdHde e7de*dede)fdQZGee dEe8j?fFddGdHde e8de*dede)fdRZHee dEe9j?fFddGdHde e9de*dede)fdSZIee dEe:j?fFddGdHde e:de*dede)fdTZJee dEe;j?fFddGdHde e;de*dede)fdUZKee dEej?fFddGdHde e>de*dede)fdXZNdS)YN) OrderedDict)partial)AnyCallableDictListOptionalTuple)nnTensor)Conv2dNormActivationSqueezeExcitation)ImageClassificationInterpolationMode)_log_api_usage_once)register_modelWeights WeightsEnum)_IMAGENET_CATEGORIES)_make_divisible_ovewrite_named_paramhandle_legacy_interface)RegNetRegNet_Y_400MF_WeightsRegNet_Y_800MF_WeightsRegNet_Y_1_6GF_WeightsRegNet_Y_3_2GF_WeightsRegNet_Y_8GF_WeightsRegNet_Y_16GF_WeightsRegNet_Y_32GF_WeightsRegNet_Y_128GF_WeightsRegNet_X_400MF_WeightsRegNet_X_800MF_WeightsRegNet_X_1_6GF_WeightsRegNet_X_3_2GF_WeightsRegNet_X_8GF_WeightsRegNet_X_16GF_WeightsRegNet_X_32GF_Weightsregnet_y_400mfregnet_y_800mfregnet_y_1_6gfregnet_y_3_2gf regnet_y_8gf regnet_y_16gf regnet_y_32gfregnet_y_128gfregnet_x_400mfregnet_x_800mfregnet_x_1_6gfregnet_x_3_2gf regnet_x_8gf regnet_x_16gf regnet_x_32gfc leZdZdZdedededejfdedejfddf fd ZxZ S) SimpleStemINz(Simple stem for ImageNet: 3x3, BN, ReLU.width_in width_out norm_layer.activation_layerreturnNcVt||dd||dS)Nr  kernel_sizestrider>r?)super__init__)selfr<r=r>r? __class__s JC:\Projects\Historical\venv\Lib\site-packages\torchvision/models/regnet.pyrGzSimpleStemIN.__init__7s>  iQqZbr      ) __name__ __module__ __qualname____doc__intrr ModulerG __classcell__rIs@rJr;r;4s22      S")^,  #3 >2             rKr;ceZdZdZdededededejfdedejfded ed e ed d ffd Z xZ S)BottleneckTransformz/Bottleneck transformation: 1x1, 3x3 [+SE], 1x1.r<r=rEr>.r? group_widthbottleneck_multiplierse_ratior@Nc t} tt||z} | |z} t|| dd||| d<t| | d|| ||| d<|r4tt||z} t | | || d<t| |dd|d| d <t | dS) NrrCarB)rDrEgroupsr>r?b)input_channelssqueeze_channels activationsec)rrProundrrrFrG)rHr<r=rEr>r?rVrWrXlayersw_bg width_se_outrIs rJrGzBottleneckTransform.__init__Fs/:mm% $99::;; ; * cqz\l   s + !F1fv   s   uX%899::L,"!-+F4L + ! ]a   s       rK rLrMrNrOrPrr rQfloatr rGrRrSs@rJrUrUCs99#!#!#! #! S")^, #! #3 >2 #!#! %#!5/#! #!#!#!#!#!#!#!#!#!#!rKrUceZdZdZ ddedededed ejfd ed ejfd ed ed e eddffd Z de de fdZ xZ S)ResBottleneckBlockz>Residual bottleneck block: x + F(x), F = bottleneck transform.r?Nr<r=rEr>.r?rVrWrXr@c td|_||kp|dk} | rt||d||d|_t |||||||||_|d|_dS)NrrCT)inplace)rFrGprojrrUfr_) rHr<r=rEr>r?rVrWrX should_projrIs rJrGzResBottleneckBlock.__init__os  9,>&A+  ,)6jkoDI%       !    +*4888rKxc|j,||||z}n|||z}||SN)rnror_rHrqs rJforwardzResBottleneckBlock.forwardsO 9  ! tvvayy(AADFF1II Aq!!!rK)rrkN)rLrMrNrOrPrr rQrhr rGr rurRrSs@rJrjrjlsHH'*$(9999 9 S")^, 9 #3 >2 99 %95/9 999999@""F""""""""rKrjceZdZdZ ddededededed ejfd ed ejfd ed ejfd ed ede ededdffd Z xZ S)AnyStagez;AnyNet stage (sequence of blocks w/ the same output shape).Nrr<r=rEdepthblock_constructor.r>r?rVrWrX stage_indexr@c tt|D]@} || dkr|n||| dkr|nd|||| | } |d| d| | AdS)Nrrblock-)rFrGrange add_module)rHr<r=rErxryr>r?rVrWrXrzir|rIs rJrGzAnyStage.__init__s u > >A%%FF q&&a %  E OO5K55!55u = = = = > >rK)NrrgrSs@rJrwrwsEE%)>>>> >  > $CN3 >S")^,>#3 >2>> %>5/>> >>>>>>>>>>rKrwcFeZdZ ddeedeedeedeedeedeeddfd Ze dd ed ed ededededeede ddfdZ dZ e deedeedeede eeeeffdZdS) BlockParamsNdepthswidths group_widthsbottleneck_multipliersstridesrXr@cZ||_||_||_||_||_||_dSrsrrrrrrX)rHrrrrrrXs rJrGzBlockParams.__init__s4  (&<#   rKrkrxw_0w_aw_mrVrWkwargsc d} d} |dks|dks|dks |dzdkrtdtj||z|z} tjtj| |z t j|z } tjtj|tj|| z| | z } tt| }t| dgzdg| z| dgzdg| z}d|D}dt| |dd D}tj tjd t|D }| g|z}|g|z}|g|z}||||\}}||||||| S) a) Programmatically compute all the per-block settings, given the RegNet parameters. The first step is to compute the quantized linear block parameters, in log space. Key parameters are: - `w_a` is the width progression slope - `w_0` is the initial width - `w_m` is the width stepping in the log space In other terms `log(block_width) = log(w_0) + w_m * block_capacity`, with `bock_capacity` ramping up following the w_0 and w_a params. This block width is finally quantized to multiples of 8. The second step is to compute the parameters per stage, taking into account the skip connection and the final 1x1 convolutions. We use the fact that the output width is constant within a stage. r rrzInvalid RegNet settingsc0g|]\}}}}||kp||kSr).0wwprrps rJ z0BlockParams.from_init_params..s-FFFB2!r'$Q"WFFFrKcg|] \}}|| Srr)rrts rJrz0BlockParams.from_init_params..s!JJJdaJJJJrKNcg|] \}}|| Srr)rdrs rJrz0BlockParams.from_init_params..s"/U/U/UdaST/U/U/U/UrKr) ValueErrortorcharangerblogmathdividepowrPtolistlensetzipdifftensor enumerate"_adjust_widths_groups_compatibilty)clsrxrrrrVrWrXrQUANTSTRIDE widths_contblock_capacity block_widths num_stages split_helpersplits stage_widths stage_depthsrrrs rJfrom_init_paramszBlockParams.from_init_paramss@ 77cQhh#((cAgll677 7l5))C/#5 UY{S/@%A%ADHSMM%QRR ELuyn7U7U1UW\$]$]^^affkkmmttvv \**++  A3  C,  A3  C,    GFFFFJJc,ss &D&DJJJ z%,/U/Ui>O>O/U/U/U"V"VWW[[]]ddff (Z'"7!8:!E#}z1 &)%K%K 0,& & " ls%#9     rKcZt|j|j|j|j|jSrs)rrrrrr)rHs rJ_get_expanded_paramsz BlockParams._get_expanded_paramss$4; dk4;LdNijjjrKrbottleneck_ratioscdt||D}dt||D}dt||D}dt||D}||fS)zl Adjusts the compatibility of widths and groups, depending on the bottleneck ratio. c8g|]\}}t||zSrrP)rrr\s rJrzBBlockParams._adjust_widths_groups_compatibilty..s&NNNA#a!e**NNNrKc4g|]\}}t||Sr)min)rrew_bots rJrzBBlockParams._adjust_widths_groups_compatibilty.. s$TTThaC5MMTTTrKc4g|]\}}t||Sr)r)rrres rJrzBBlockParams._adjust_widths_groups_compatibilty..#s&ZZZq/%++ZZZrKc8g|]\}}t||z Srr)rrr\s rJrzBBlockParams._adjust_widths_groups_compatibilty..$s&VVV85!EAIVVVrK)r)rrrrgroup_widths_minws_bots rJrz.BlockParams._adjust_widths_groups_compatibiltysON\;L)M)MNNNTT#lF:S:STTT[ZCHXr_r@c2tt||t}| tj}|t }| tj}|d||||_|}g} t| D]J\} \} } } }}| d| dzt|| | | ||||||j | dz f| }Kt jt| |_t jd|_t j|||_|D]E}t-|tjr`|jd|jdz|jz}tj|jdt;jd |z  }t-|tjrItj|jtj |j!t-|tjrKtj|jdd tj |j!GdS) NrBr|r)rzrr) in_features out_featuresrg@)meanstdg{Gz?)"rFrGrr;r BatchNorm2drjReLUstemrrappendrwrX Sequentialr trunk_outputAdaptiveAvgPool2davgpoolLinearfcmodules isinstanceConv2drD out_channelsinitnormal_weightrsqrtones_zeros_bias)rHrrrrrr>r_ current_widthblocksrr=rErxrVrWmfan_outrIs rJrGzRegNet.__init__)s~ D!!!  $I  J  +J  JI       # |88:: ; ;  & & A     ! MM!AaCMM%!"""#-$-$%E      &&MMM+f*=*=>>+F33 ) KPPP ' 'A!RY'' '-*Q]1-==Ns #-8P8PQQQQAr~.. ' ah'''qv&&&&Ary)) 's===qv&&& ' 'rKrqc||}||}||}|d}||}|S)Nr) start_dim)rrrflattenrrts rJruzRegNet.forwardxsZ IIaLL   a  LLOO III " " GGAJJrK)rrNNNN) rLrMrNrrPr rr rQrGr rurRrSs@rJrr(s 8<9=9=9=M'M'!M'M' M' HS")^45 M' Xc29n56 M'Xc29n56M'Xc29n56M' M'M'M'M'M'M'^FrKrrweightsprogressrr@c 2|)t|dt|jd|dt t jdd}t|fd|i|}|*|| |d|S) Nr categoriesr>gh㈵>g?)epsmomentumT)r check_hash) rrmetapoprr rrload_state_dictget_state_dict)rrrrr>models rJ_regnetrs fmSl9S5T5TUUUL'".eVY*Z*Z*Z[[J < A AJ A& A AE g44hSW4XXYYY LrKr)min_sizer _COMMON_METAz(https://github.com/facebookresearch/SWAGz:https://github.com/facebookresearch/SWAG/blob/main/LICENSE)recipelicensec eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)rz?https://download.pytorch.org/models/regnet_y_400mf-c65dace8.pth crop_sizeiPIBRhttps://github.com/pytorch/vision/tree/main/references/classification#small-models ImageNet-1KgmR@gV@zacc@1zacc@5g|?5^?guV0@XThese weights reproduce closely the results of the paper using a simple training recipe. num_paramsr_metrics_ops _file_size_docsurl transformsrz?https://download.pytorch.org/models/regnet_y_400mf-e6988f5f.pthr resize_size8https://github.com/pytorch/vision/issues/3995#new-recipeg~jtR@gsh|/W@$ These weights improve upon the results of the original paper by using a modified version of TorchVision's `new training recipe `_. N rLrMrNrrrr IMAGENET1K_V1 IMAGENET1K_V2DEFAULTrrKrJrrG M7.#>>>    !j##    s    M$G M7.#3OOO  !P##       M,GGGrKrc eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)rz?https://download.pytorch.org/models/regnet_y_800mf-1b27b58c.pthrri'brrg{GS@gbX9HW@r g㥛 ?g/$8@r r rz?https://download.pytorch.org/models/regnet_y_800mf-58fc7688.pthrrrgES@g㥛 W@rNrrrKrJrrrrKrc eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)rz?https://download.pytorch.org/models/regnet_y_1_6gf-b11a554e.pthrri~rrg|S@g}W@r gn?g~jtE@r r rz?https://download.pytorch.org/models/regnet_y_1_6gf-0d7bc02a.pthrrrgMb8T@g#~jW@rNrrrKrJrrsG M7.#>>>    "j##    s    M$G M7.#3OOO  "P##       M,GGGrKrc eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)rz?https://download.pytorch.org/models/regnet_y_3_2gf-b5a9779c.pthrri2(Shttps://github.com/pytorch/vision/tree/main/references/classification#medium-modelsrg&1S@g/ݤW@r g rh @g?5^IR@r r rz?https://download.pytorch.org/models/regnet_y_3_2gf-9180c971.pthrrrg+~T@g|?5W@rNrrrKrJrr%G M7.#>>>    "k##    s    M$G M7.#3OOO  "P##       M,GGGrKrc eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)r z=https://download.pytorch.org/models/regnet_y_8gf-d0d0e4a8.pthrriXr!rg5^I T@gPnW@r gV- @gOnb@r r rz=https://download.pytorch.org/models/regnet_y_8gf-dc2b1b54.pthrrrgET@gQX@rNrrrKrJr r QG K7.#>>>    "k##   !s    M$G K7.#3OOO  "P##   !   M,GGGrKr c veZdZedeediedddddd id d d d Zedeeddiedddddd id d dd Zedeedde j ie ddddd idd ddZ edeedde j ie dddd d!d id d d"d#Z eZd$S)%r!z>https://download.pytorch.org/models/regnet_y_16gf-9e6ed7dd.pthrri{Rhttps://github.com/pytorch/vision/tree/main/references/classification#large-modelsrgB`"T@g(\W@r gm/@gp= s@r r rz>https://download.pytorch.org/models/regnet_y_16gf-3e4a00f9.pthrrrgbX9T@gEX@rzChttps://download.pytorch.org/models/regnet_y_16gf_swag-43afe44d.pthrr interpolationgT㥛ĀU@g~jtX@gGz^G@ These weights are learnt via transfer learning by end-to-end fine-tuning the original `SWAG `_ weights on ImageNet-1K data. r r rrrzFhttps://download.pytorch.org/models/regnet_y_16gf_lc_swag-f3ec0043.pth+https://github.com/pytorch/vision/pull/5793gX9vT@gV-OX@ These weights are composed of the original frozen `SWAG `_ trunk weights and a linear classifier learnt on top of them trained on ImageNet-1K data. rr r rrrNrLrMrNrrrrrrrBICUBIC_COMMON_SWAG_METAIMAGENET1K_SWAG_E2E_V1IMAGENET1K_SWAG_LINEAR_V1rrrKrJr!r!}sG L7.#>>>    "j##    s    M$G L7.#3OOO  "P##       M,%W Q7 3CO`Oh     "##       ,!( T7 3CO`Oh     C"##       !!!.GGGrKr!c veZdZedeediedddddd id d d d Zedeeddiedddddd id d dd Zedeedde j ie ddddd idd ddZ edeedde j ie dddd d!d id d d"d#Z eZd$S)%r"z>https://download.pytorch.org/models/regnet_y_32gf-4dee3f7a.pthrri<r&rgx&18T@g(\W@r gp= #@@gS㥛P@r r rz>https://download.pytorch.org/models/regnet_y_32gf-8db6d4b5.pthrrrgd;OT@gZd;X@rzChttps://download.pytorch.org/models/regnet_y_32gf_swag-04fdfa75.pthr'r(gʡU@gI +X@g/ݴW@r*r+zFhttps://download.pytorch.org/models/regnet_y_32gf_lc_swag-e1583746.pthr,g+'U@gQ^X@r-r.Nr/rrKrJr"r"sG L7.#>>>    #j##   !s    M$G L7.#3OOO  #P##   !   M,%W Q7 3CO`Oh     ###   !   ,!( T7 3CO`Oh     C###   !   !!!.GGGrKr"c eZdZedeeddejiedddddid d d d  Z edeeddejieddddddidd dd Z e Z dS)r#zDhttps://download.pytorch.org/models/regnet_y_128gf_swag-c8ce3e52.pthr'r(i^o&rg;OV@gS㥫X@r gQiw@g㥛 ;@r*r+rzGhttps://download.pytorch.org/models/regnet_y_128gf_lc_swag-cbe8ce12.pthrr,g1ZU@gtvX@gx&_@r-r.N) rLrMrNrrrrr0r1r2r3rrrKrJr#r#/s$W R7 3CO`Oh     ###   "   ,!( U7 3CO`Oh     C###   "   !!!.%GGGrKr#c eZdZedeediedddddd id d d d Zedeeddiedddddd id ddd ZeZ dS)r$z?https://download.pytorch.org/models/regnet_x_400mf-adf1edd5.pthrriSrrgL7A`5R@g̼V@r g"~?g5^I B5@r r rz?https://download.pytorch.org/models/regnet_x_400mf-62229a5f.pthrrDhttps://github.com/pytorch/vision/issues/3995#new-recipe-with-fixresgKR@gS㥛W@goA5@rNrrrKrJr$r$`G M7.#>>>    !j##    s    M$G M7.#3OOO  !\##       M,GGGrKr$c eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)r%z?https://download.pytorch.org/models/regnet_x_800mf-ad17e45c.pthrrinrrg!rhR@gʡEW@r g?gR;@r r rz?https://download.pytorch.org/models/regnet_x_800mf-94a99ebd.pthrrr7g rhaS@g/tW@rNrrrKrJr%r%sG M7.#>>>    !j##    s    M$G M7.#3OOO  !\##       M,GGGrKr%c eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)r&z?https://download.pytorch.org/models/regnet_x_1_6gf-e3633e7f.pthrri:rrg(\BS@g\(\W@r gS?gZdA@r r rz?https://download.pytorch.org/models/regnet_x_1_6gf-a12f2b72.pthrrr7gnS@g^I W@rNrrrKrJr&r&r8rKr&c eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)r'z?https://download.pytorch.org/models/regnet_x_3_2gf-f342aeae.pthrri(hr!rgKS@gsh|W@r g"~j @gT㥛`M@r r rz?https://download.pytorch.org/models/regnet_x_3_2gf-7071aa85.pthrrrgCLT@gQW@rNrrrKrJr'r'r"rKr'c eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)r(z=https://download.pytorch.org/models/regnet_x_8gf-03ceed89.pthrri[r!rgtS@gClW@r g{G@g;Ob@r r rz=https://download.pytorch.org/models/regnet_x_8gf-2b70d774.pthrrrgSkT@gZdW@rNrrrKrJr(r(r$rKr(c eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)r)z>https://download.pytorch.org/models/regnet_x_16gf-2007eb11.pthrri9<r!rgʡET@g#~jW@r go/@gMbi@r r rz>https://download.pytorch.org/models/regnet_x_16gf-ba3796d7.pthrrrgҭT@gC X@rNrrrKrJr)r)<sG L7.#>>>    "k##   !s    M$G L7.#3OOO  "P##   !   M,GGGrKr)c eZdZedeediedddddd id d d d Zedeeddiedddddd id d dd ZeZ dS)r*z>https://download.pytorch.org/models/regnet_x_32gf-9d47f8d0.pthrrimr&rg+'T@gZd;W@r g#~j?@gvy@r r rz>https://download.pytorch.org/models/regnet_x_32gf-6eb8fdc6.pthrrrg7A`T@gOnX@rNrrrKrJr*r*hsG L7.#>>>    #j##   !s    M$G L7.#3OOO  #P##   !   M,GGGrKr* pretrained)rT)rrc t|}tjdddddddd|}t |||fi|S) a Constructs a RegNetY_400MF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_Y_400MF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_Y_400MF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_Y_400MF_Weights :members: 0gp= ;@gQ@r?rxrrrrVrXr)rverifyrrrrrrparamss rJr+r+sY(%++G44G  ) x4]^im x xqw x xF 67H 7 7 7 77rKc t|}tjdddddddd|}t |||fi|S) a Constructs a RegNetY_800MF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_Y_800MF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_Y_800MF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_Y_800MF_Weights :members: 8gQkC@g333333@rArCrDr)rrErrrrFs rJr,r,sY(%++G44G  ) x3\^im x xqw x xF 67H 7 7 7 77rKc t|}tjdddddddd|}t |||fi|S) a Constructs a RegNetY_1.6GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_Y_1_6GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_Y_1_6GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_Y_1_6GF_Weights :members: rBg(\µ4@g333333@rCrDr)rrErrrrFs rJr-r-d(%++G44G  )be2PVF 67H 7 7 7 77rKc t|}tjdddddddd|}t |||fi|S) a Constructs a RegNetY_3.2GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_Y_3_2GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_Y_3_2GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_Y_3_2GF_Weights :members: Pgq= ףPE@gHzG@rMrCrDr)rrErrrrFs rJr.r.rNrKc t|}tjdddddddd|}t |||fi|S) a Constructs a RegNetY_8GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_Y_8GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_Y_8GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_Y_8GF_Weights :members: gGz4S@gQ@rJrCrDr)r rErrrrFs rJr/r/sd(#))'22G  )cu$BQWF 67H 7 7 7 77rKc t|}tjdddddddd|}t |||fi|S) a Constructs a RegNetY_16GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_Y_16GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_Y_16GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_Y_16GF_Weights :members: gQZ@gףp= @prCrDr)r!rErrrrFs rJr0r0d($**733G  )cv4S4SYF 67H 7 7 7 77rKc t|}tjdddddddd|}t |||fi|S)a Constructs a RegNetY_32GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_Y_32GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_Y_32GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_Y_32GF_Weights :members: rg)\\@g= ףp=@rCrDr)r"rErrrrFs rJr1r18rYrK)r?Nc t|}tjdddddddd|}t |||fi|S) a Constructs a RegNetY_128GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_Y_128GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_Y_128GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_Y_128GF_Weights :members: rLig(\d@g)\(@irCrDr)r#rErrrrFs rJr2r2Tsd(%++G44G  )cv4S4SYF 67H 7 7 7 77rKc t|}tjddddddd|}t |||fi|S)a Constructs a RegNetX_400MF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_X_400MF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_X_400MF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_X_400MF_Weights :members: rMg{Gz8@gRQ@rArxrrrrVr)r$rErrrrFs rJr3r3pV(%++G44G  ) j4]_ j jci j jF 67H 7 7 7 77rKc t|}tjddddddd|}t |||fi|S)a Constructs a RegNetX_800MF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_X_800MF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_X_800MF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_X_800MF_Weights :members: rArJg= ףpA@g= ףp=@r_r)r%rErrrrFs rJr4r4r`rKc t|}tjddddddd|}t |||fi|S)a Constructs a RegNetX_1.6GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_X_1_6GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_X_1_6GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_X_1_6GF_Weights :members: rVrQgzGA@@rMr_r)r&rErrrrFs rJr5r5r`rKc t|}tjddddddd|}t |||fi|S)a Constructs a RegNetX_3.2GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_X_3_2GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_X_3_2GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_X_3_2GF_Weights :members: Xg(\O:@rcrBr_r)r'rErrrrFs rJr6r6r`rKc t|}tjddddddd|}t |||fi|S)a Constructs a RegNetX_8GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_X_8GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_X_8GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_X_8GF_Weights :members: rQgHzH@g ףp= @xr_r)r(rErrrrFs rJr7r7sV(#))'22G  ) k4]` k kdj k kF 67H 7 7 7 77rKc t|}tjddddddd|}t |||fi|S)a Constructs a RegNetX_16GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_X_16GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_X_16GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_X_16GF_Weights :members: r^gQK@g@r_r)r)rErrrrFs rJr8r8V($**733G  ) kC]` k kdj k kF 67H 7 7 7 77rKc t|}tjddddddd|}t |||fi|S)a Constructs a RegNetX_32GF architecture from `Designing Network Design Spaces `_. Args: weights (:class:`~torchvision.models.RegNet_X_32GF_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.RegNet_X_32GF_Weights` below for more details and possible values. By default, no pretrained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to either ``torchvision.models.regnet.RegNet`` or ``torchvision.models.regnet.BlockParams`` class. Please refer to the `source code `_ for more detail about the classes. .. autoclass:: torchvision.models.RegNet_X_32GF_Weights :members: rhi@gףp= wQ@rr_r)r*rErrrrFs rJr9r9 rmrK)Or collectionsr functoolsrtypingrrrrr r rr r ops.miscrrtransforms._presetsrrutilsr_apirrr_metar_utilsrrr__all__r;rrUrQrjrwrrboolrrstr__annotations__r1rrrrr r!r"r#r$r%r&r'r(r)r*rr+r,r-r.r/r0r1r2r3r4r5r6r7r8r9rrKrJr}s  ######================ >>>>>>>>HHHHHHHH''''''6666666666''''''SSSSSSSSSS   F      '    &!&!&!&!&!"-&!&!&!R("("("("("("("("V>>>>>r}>>>Dl.l.l.l.l.l.l.l.^XXXXXRYXXXv k "   &&   d38n 8K)))))[)))X)))))[)))X)))))[)))X)))))[)))X)))));)))XVVVVVKVVVrVVVVVKVVVr.%.%.%.%.%[.%.%.%b)))))[)))X)))))[)))X)))))[)))X)))))[)))X)))));)))X)))))K)))X)))))K)))X,0F0T!UVVVBFY]888x(>?8RV8il8qw888WV80,0F0T!UVVVBFY]888x(>?8RV8il8qw888WV80,0F0T!UVVVBFY]888x(>?8RV8il8qw888WV84,0F0T!UVVVBFY]888x(>?8RV8il8qw888WV84,0D0R!STTT>BUY888X&:;8d8eh8ms888UT84,0E0S!TUUU@DW[888h'<=8PT8gj8ou888VU84,0E0S!TUUU@DW[888h'<=8PT8gj8ou888VU84!5666BFY]888x(>?8RV8il8qw8887684,0F0T!UVVVBFY]888x(>?8RV8il8qw888WV80,0F0T!UVVVBFY]888x(>?8RV8il8qw888WV80,0F0T!UVVVBFY]888x(>?8RV8il8qw888WV80,0F0T!UVVVBFY]888x(>?8RV8il8qw888WV80,0D0R!STTT>BUY888X&:;8d8eh8ms888UT80,0E0S!TUUU@DW[888h'<=8PT8gj8ou888VU80,0E0S!TUUU@DW[888h'<=8PT8gj8ou888VU888rK