+ j;<^RIHt^RIHt^RIHtHtHt^RIt^RIH t H t ^RI H t ^RIHtHt^RIHt^R IHt^R IHt^R IHtHtHt^R IHt^R IHtHt.R3Ot !RR] PB4t"!RR] PF4t$!RR4t%!RR] PF4t&RRlt'RR4R]RRR R!/t(!R"R]4t)!R#R]4t*!R$R]4t+!R%R]4t,]!4]!R&])PZ3R'7R(RR)R*/R+R,ll44t.]!4]!R&]*PZ3R'7R(RR)R*/R-R.ll44t/]!4]!R&]+PZ3R'7R(RR)R*/R/R0ll44t0]!4]!R&],PZ3R'7R(RR)R*/R1R2ll44t1R#)5)Sequence)partial)AnyCallableOptionalN)nnTensor) functional)Conv2dNormActivationPermute)StochasticDepth)ImageClassification)_log_api_usage_once)register_modelWeights WeightsEnum)_IMAGENET_CATEGORIES)_ovewrite_named_paramhandle_legacy_interfaceConvNeXtConvNeXt_Tiny_WeightsConvNeXt_Small_WeightsConvNeXt_Base_WeightsConvNeXt_Large_Weightsc2a]tRt^toV3RlRltRtVtR#) LayerNorm2dc&<V^8dQhRS[RS[/#xreturnr )format __classdict__s"\/home/luca/screen-checking/.venv/lib/python3.14/site-packages/torchvision/models/convnext.py __annotate__LayerNorm2d.__annotate__ sFcVP^^^^4p\P!WPVPVP VP 4pVP^^^^4pV#)r)permuteF layer_normnormalized_shapeweightbiasepsselfr s&&r%forwardLayerNorm2d.forward sU IIaAq ! LL114;; 488 T IIaAq !r(N)__name__ __module__ __qualname____firstlineno__r3__static_attributes____classdictcell__r$s@r%rrsr(rcTaa]tRt^'toRV3RlV3RllltV3RlRltRtVtV;t#)CNBlockc h<V^8dQhRS[RS[RS[S[RS[P3,,RR/#)r layer_scalestochastic_depth_prob norm_layer.r!N)floatrrrModule)r#r$s"r%r&CNBlock.__annotate__(sMNNN % N Xc299n56 N  Nr(c ><\SV`4Vf\\PRR7p\P !\P !W^^VRR7\.RO4V!V4\P!V^V,RR7\P!4\P!^V,VRR7\.RO44Vn \P!\P!V^^4V,4Vn\VR4VnR#) Nư>r0T) kernel_sizepaddinggroupsr/) in_features out_featuresr/row)rr)rrOrPr)super__init__rr LayerNorm SequentialConv2dr LinearGELUblock Parametertorchonesr@r stochastic_depth)r2dimr@rArB __class__s&&&&&r%rRCNBlock.__init__(s    48J]] IIcAq4 P L ! sO II#AG$ G GGI II!c'$ G L !  << 31(= (KL /0Eu Mr(c&<V^8dQhRS[RS[/#)rinputr!r")r#r$s"r%r&rE?sVr(c|VPVPV4,pVPV4pW!, pV#N)r@rXr\)r2raresults&& r%r3CNBlock.forward?s7!!DJJu$55&&v. r()rXr@r\rc) r6r7r8r9rRr3r:r; __classcell__r^r$s@@r%r>r>'s NN.r(r>cDa]tRt^FtoV3RlRltV3RlRltRtVtR#) CNBlockConfigc@<V^8dQhRS[RS[S[,RS[RR/#)rinput_channels out_channels num_layersr!N)intr)r#r$s"r%r&CNBlockConfig.__annotate__Hs7%%%sm% %  %r(c*WnW nW0nR#rc)rkrlrm)r2rkrlrms&&&&r%rRCNBlockConfig.__init__Hs -($r(c <V^8dQhRS[/#)rr!)str)r#r$s"r%r&roRs))#)r(cVPPR,pVR, pVR, pVR, pVR, pVP!R/VPB#)(zinput_channels={input_channels}z, out_channels={out_channels}z, num_layers={num_layers})r5)r^r6r#__dict__)r2ss& r%__repr__CNBlockConfig.__repr__RsV NN # #c ) .. ,, (( Sxx($--((r()rkrmrlN)r6r7r8r9rRryr:r;r<s@r%ririFs%%))r(ricfaa]tRt^[toRV3RlV3RllltV3RlRltV3RlRltRtVtV;t #) rc<V^8dQhRS[S[,RS[RS[RS[RS[S[RS[P3,,RS[S[RS[P3,,RS[R R /#) r block_settingrAr@ num_classesrX.rBkwargsr!N) listrirCrnrrrrDr)r#r$s"r%r&ConvNeXt.__annotate__\sL+L+M*L+ %L+ L+  L+ bii01 L+Xc299n56L+L+ L+r(c <\SV`4\V4V'g \R4h\ V\ 4'd0\ VUu.uFp\ V\4NK up4'g \R4hVf\pVf\\RR7p.p V^,Pp V P\^V ^^^VRRR74\RV44p ^p VFp .p\!V P"4FEpW,,V R, , pVPV!V PVV44V ^, p KG V P\$P&!V!4V P(fKV P\$P&!V!V P4\$P*!V PV P(^^R 744K \$P&!V !Vn\$P.!^4VnVR ,pVP(e VP(M VPp\$P&!V!V4\$P2!^4\$P4!VV44VnVP94Fp\ V\$P*\$P434'gK5\$P:P=VP>R R 7VP@fKp\$P:PCVP@4K R#uupi) z%The block_setting should not be emptyz/The block_setting should be List[CNBlockConfig]NrGrHT)rIstriderJrBactivation_layerr/c38"TFqPxK R#5irc)rm).0cnfs& r% $ConvNeXt.__init__..s I=C=sg?)rIrg{Gz?)std)"rQrRr ValueError isinstancerallri TypeErrorr>rrrkappendr sumrangermrrTrlrUfeaturesAdaptiveAvgPool2davgpoolFlattenrV classifiermodulesinit trunc_normal_r.r/zeros_)r2r}rAr@r~rXrBrrxlayersfirstconv_output_channelstotal_stage_blocksstage_block_idrstage_sd_prob lastblocklastconv_output_channelsmr^s&&&&&&&, r%rRConvNeXt.__init__\s D!DE E]H55#er>ser`az!]?[er>s:t:tMN N =E   $7J"$%2!$4$C$C! )%!%  ! I= II C%'E3>>*/@DVY\D\] U3#5#5{GLM!# + MM"--/ 0+ MM"3#5#56 #"4"4c6F6FTU^_`!$ v. ++A. !"% &/&<&<&HI " "iNfNf !-- / 0"**Q-KcepAq A!bii344%%ahhD%966%GGNN166*  s?ts Mc&<V^8dQhRS[RS[/#rr")r#r$s"r%r&rsv&r(clVPV4pVPV4pVPV4pV#rc)rrrr1s&&r% _forward_implConvNeXt._forward_impls0 MM!  LLO OOA r(c&<V^8dQhRS[RS[/#rr")r#r$s"r%r&rs%%%F%r(c$VPV4#rc)rr1s&&r%r3ConvNeXt.forwards!!!$$r()rrr)grGiNN) r6r7r8r9rRrr3r:r;rfrgs@@r%rr[s+L+L+\ %%%r(c V^8dQhR\\,R\R\\,R\ R\ R\/#)rr}rAweightsprogressrr!)rrirCrrboolrr)r#s"r%r&r&sN & k "     r(c Ve(\VR\VPR,44\V3RV/VBpVe#VP VP VRR74V#)Nr~ categoriesrAT)r check_hash)rlenmetarload_state_dictget_state_dict)r}rArrrmodels&&&&, r% _convnextrsdfmSl9S5TU ] Z:O ZSY ZE g44hSW4XY Lr(min_sizerrecipezNhttps://github.com/pytorch/vision/tree/main/references/classification#convnext_docsa These weights improve upon the results of the original paper by using a modified version of TorchVision's `new training recipe `_. c f]tRt^t]!R]!]^^R7/]CRRRRRRR R //R R R R/CR7t]t Rt R#)rz>https://download.pytorch.org/models/convnext_tiny-983f1562.pth crop_size resize_size num_paramsiH<_metrics ImageNet-1Kacc@1gzGT@acc@5gMbX X@_opsgm@ _file_sizegV-G[@url transformsrr5N r6r7r8r9rrr _COMMON_META IMAGENET1K_V1DEFAULTr:r5r(r%rrs` L.#3O    ( VV  E '  M Gr(c f]tRt^t]!R]!]^^R7/]CRRRRRRR R //R R R R/CR7t]t Rt R#)rz?https://download.pytorch.org/models/convnext_small-0c510722.pthrriHZrrrgClT@rg)X@rg|?5^!@rg"~g@rr5Nrr5r(r%rrs` M.#3O    ( VV  E '  M Gr(c f]tRt^t]!R]!]^^R7/]CRRRRRRR R //R R R R/CR7t]t Rt R#)rz>https://download.pytorch.org/models/convnext_base-6075fbad.pthrrihGrrrgU@rgHz7X@rg(\µ.@rg/$!u@rr5Nrr5r(r%rrs` L.#3O    ( VV  F '  M Gr(c f]tRtRt]!R]!]^^R7/]CRRRRRR R R //R R RR/CR7t]t Rt R#)riz?https://download.pytorch.org/models/convnext_large-ea097f82.pthrri rrrg"~U@rgX9v>X@rg|?5.A@rgK@rr5Nrr5r(r%rrs` M.#3O    ) VV  F '  M Gr( pretrained)rrrTc^V^8dQhR\\,R\R\R\/#rrrrr!)rrrrr)r#s"r%r&r&$7XXh'<=XPTXgjXowXr(c \PV4p\^`^^4\^R^4\RR^ 4\RR^4.pVPRR4p\ W4W3/VB#)aMConvNeXt Tiny model architecture from the `A ConvNet for the 2020s `_ paper. Args: weights (:class:`~torchvision.models.convnext.ConvNeXt_Tiny_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.convnext.ConvNeXt_Tiny_Weights` below for more details and possible values. By default, no pre-trained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to the ``torchvision.models.convnext.ConvNext`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.ConvNeXt_Tiny_Weights :members: NrAg?)rverifyripoprrrrr}rAs$$, r% convnext_tinyr"st&$**73G b#q!c3"c3"c4# M #JJ'>D ]7 WPV WWr(c^V^8dQhR\\,R\R\R\/#r)rrrrr)r#s"r%r&r&C8XX/0XCGXZ]X Xr(c \PV4p\^`^^4\^R^4\RR^4\RR^4.pVPRR4p\ W4W3/VB#)aQConvNeXt Small model architecture from the `A ConvNet for the 2020s `_ paper. Args: weights (:class:`~torchvision.models.convnext.ConvNeXt_Small_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.convnext.ConvNeXt_Small_Weights` below for more details and possible values. By default, no pre-trained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to the ``torchvision.models.convnext.ConvNext`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.ConvNeXt_Small_Weights :members: rrNrAg?)rrrirrrs$$, r%convnext_smallrAst*%++G4G b#q!c3"c3#c4# M #JJ'>D ]7 WPV WWr(c^V^8dQhR\\,R\R\R\/#r)rrrrr)r#s"r%r&r&drr(c \PV4p\^R^4\RR^4\RR^4\RR^4.pVPRR4p\ W4W3/VB#)aMConvNeXt Base model architecture from the `A ConvNet for the 2020s `_ paper. Args: weights (:class:`~torchvision.models.convnext.ConvNeXt_Base_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.convnext.ConvNeXt_Base_Weights` below for more details and possible values. By default, no pre-trained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to the ``torchvision.models.convnext.ConvNext`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.ConvNeXt_Base_Weights :members: iiNrA?)rrrirrrs$$, r% convnext_baserbst&$**73G c3"c3"c4$dD!$ M #JJ'>D ]7 WPV WWr(c^V^8dQhR\\,R\R\R\/#r)rrrrr)r#s"r%r&r&rr(c \PV4p\^R^4\RR^4\RR^4\RR^4.pVPRR4p\ W4W3/VB#)aQConvNeXt Large model architecture from the `A ConvNet for the 2020s `_ paper. Args: weights (:class:`~torchvision.models.convnext.ConvNeXt_Large_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.convnext.ConvNeXt_Large_Weights` below for more details and possible values. By default, no pre-trained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to the ``torchvision.models.convnext.ConvNext`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.ConvNeXt_Large_Weights :members: rriNrAr)rrrirrrs$$, r%convnext_largerst*%++G4G c3"c3"c4$dD!$ M #JJ'>D ]7 WPV WWr() rrrrrrrrr) r)2collections.abcr functoolsrtypingrrrrZrr torch.nnr r+ops.miscr r ops.stochastic_depthr transforms._presetsrutilsr_apirrr_metar_utilsrr__all__rSrrDr>rirrrrrrrrrrrrr5r(r%rs$** $425'66'B ",,bii>))*V%ryyV%r&& ^   K([(K([(,0E0S0S!TUXXW[XVX:,0F0T0T!UVX37XJNXWX>,0E0S0S!TUXXW[XVX:,0F0T0T!UVX37XJNXWXr(