cCiOSrSSKrSSKJrJr SSKrSSKJr SSKJr SSK J r J r J r SSK Jr SS KJrJr S S KJr \R("\5rS \S \4SjrS+S\\\4S\4Sjjr"SS\R85r"SS\R<5r"SS\R85r "SS\R85r!"SS\R85r""SS\R85r#"SS\R85r$"S S!\R85r%\"S"S#\55r&\"S$S%\&55r'\"S&S'9"S(S)\&55r(/S*Qr)g),zPyTorch EfficientNet model.N)OptionalUnion)nn)ACT2FN)BaseModelOutputWithNoAttention(BaseModelOutputWithPoolingAndNoAttention$ImageClassifierOutputWithNoAttention)PreTrainedModel)auto_docstringlogging)EfficientNetConfigconfig num_channelscURnXR-n[U[XS- -5U-U-5nUSU-:aX2- n[U5$)z4 Round number of filters based on depth multiplier. g?) depth_divisorwidth_coefficientmaxint)rrdivisornew_dims p/home/james-whalen/.local/lib/python3.13/site-packages/transformers/models/efficientnet/modeling_efficientnet.py round_filtersr%s`""G,,,L'3|k9:gEOPG|## w< kernel_sizeadjustc[U[5(aX4nUSS-USS-4nU(aUSS- USUSS- US4$USUSUSUS4$)a. Utility function to get the tuple padding value for the depthwise convolution. Args: kernel_size (`int` or `tuple`): Kernel size of the convolution layers. adjust (`bool`, *optional*, defaults to `True`): Adjusts padding value to apply to right and bottom sides of the input. rrr) isinstancer)rrcorrects r correct_padr"4s~+s##"0 1~"KNa$78G  Q GAJNGAJGG GAJ GAJ??rcn^\rSrSrSrS\4U4SjjrS\RS\R4Sjr Sr U=r $) EfficientNetEmbeddingsHzD A module that corresponds to the stem module of the original work. rc |>[TU]5 [US5Ul[R "SS9Ul[R"URURSSSSS9Ul [R"URURURS 9Ul [URUlg) N )rrrrpaddingrrvalidFrstrider)bias)epsmomentum)super__init__rout_dimr ZeroPad2dr)Conv2dr convolution BatchNorm2dbatch_norm_epsbatch_norm_momentum batchnormr hidden_act activationselfr __class__s rr1EfficientNetEmbeddings.__init__Ms $VR0 ||L9 99   1QPW^c  &:O:OZ`ZtZtu !2!23r pixel_valuesreturncURU5nURU5nURU5nURU5nU$N)r)r5r9r;)r=r@featuress rforwardEfficientNetEmbeddings.forwardXsA<< -##H->>(+??8,r)r;r9r5r2r)) __name__ __module__ __qualname____firstlineno____doc__rr1torchTensorrE__static_attributes__ __classcell__r>s@rr$r$Hs5 41 4ELLU\\rr$c:^\rSrSrSU4SjjrSrU=r$)EfficientNetDepthwiseConv2dac 8>X-n [T U]UU UUUUUUUS9 g)N) in_channels out_channelsrr,r)dilationgroupsr- padding_mode)r0r1) r=rUdepth_multiplierrr,r)rWr-rYrVr>s rr1$EfficientNetDepthwiseConv2d.__init__bs:#5  #%#%  r)rrrrrTzeros)rGrHrIrJr1rNrOrPs@rrRrRas$   rrRcz^\rSrSrSrS\S\S\S\4U4SjjrS\RS \R4S jr S r U=r $) EfficientNetExpansionLayer{zW This corresponds to the expansion phase of each block in the original implementation. rin_dimr2r,c>[TU]5 [R"UUSSSS9Ul[R "X1R S9Ul[URUl g)NrsameFrUrVrr)r-) num_featuresr.) r0r1rr4 expand_convr6r7 expand_bnrr: expand_act)r=rrar2r,r>s rr1#EfficientNetExpansionLayer.__init__sX 99   WBWBWX !2!23r hidden_statesrAclURU5nURU5nURU5nU$rC)rfrgrhr=rjs rrE"EfficientNetExpansionLayer.forwards4((7 }5  6 r)rhrgrf)rGrHrIrJrKrrr1rL FloatTensorrMrErNrOrPs@rr_r_{sM 41 43 4 4VY 4U%6%65<<rr_c ~^\rSrSrSrS\S\S\S\S\4 U4SjjrS \ RS \ R4S jr S r U=r$) EfficientNetDepthwiseLayerzc This corresponds to the depthwise convolution phase of each block in the original implementation. rrar,radjust_paddingcF>[TU]5 X0lURS:XaSOSn[XES9n[R "US9Ul[X$X6SS9Ul[R"X!RURS9Ul [URUlg) Nrr*rc)rr(Fr+rer.r/)r0r1r,r"rr3depthwise_conv_padrRdepthwise_convr6r7r8depthwise_normrr: depthwise_act) r=rrar,rrrconv_padr)r>s rr1#EfficientNetDepthwiseLayer.__init__s  "kkQ.7FkA"$,,w"?9 FSX !nn%:%:VE_E_ $F$5$56rrjrAcURS:XaURU5nURU5nURU5nUR U5nU$)Nr)r,rurvrwrxrls rrE"EfficientNetDepthwiseLayer.forwardsT ;;!  33MBM++M: ++M: **=9 r)rxrvrurwr,rGrHrIrJrKrrboolr1rLrnrMrErNrOrPs@rrprps_7"77 7  7  7, U%6%6 5<<  rrpc ~^\rSrSrSrS S\S\S\S\4U4SjjjrS\ RS \ R4S jr S r U=r$) EfficientNetSqueezeExciteLayerzd This corresponds to the Squeeze and Excitement phase of each block in the original implementation. rra expand_dimexpandc>[TU]5 U(aUOUUl[S[ X!R -55Ul[R"SS9Ul [R"URUR SSS9Ul [R"UR URSSS9Ul [URUl[R "5Ulg)Nr) output_sizerc)rUrVrr))r0r1dimrrsqueeze_expansion_ratiodim_serAdaptiveAvgPool2dsqueezer4reducerrr: act_reduceSigmoid act_expand)r=rrarrr>s rr1'EfficientNetSqueezeExciteLayer.__init__s !':V!S*H*H!HIJ ++: ii   ii    !!2!23**,rrjrAcUnURU5nURU5nURU5nURU5nUR U5n[ R "X!5nU$rC)rrrrrrLmul)r=rjinputss rrE&EfficientNetSqueezeExciteLayer.forwardsa ]3  M2  6  M2  6  &8 r)rrrrrrr)Fr}rPs@rrrsR'1'3'C'Y]''* U%6%6 5<<  rrc ^\rSrSrSrS\S\S\S\S\S\4 U4S jjr S \ RS \ RS \ R4S jr SrU=r$)EfficientNetFinalBlockLayerzS This corresponds to the final phase of each block in the original implementation. rrar2r, drop_rateid_skipc>[TU]5 US:H=(a U(+Ul[R"UUSSSS9Ul[R "X1RURS9Ul [R"US9Ul g)NrrcFrdrtp) r0r1 apply_dropoutrr4 project_convr6r7r8 project_bnDropoutdropout)r=rrar2r,rrr>s rr1$EfficientNetFinalBlockLayer.__init__sx #q[8[II   .. &;&;fF`F` zzI. r embeddingsrjrAcURU5nURU5nUR(aURU5nX!-nU$rC)rrrr)r=rrjs rrE#EfficientNetFinalBlockLayer.forwardsE))-8  6    LL7M)6Mr)rrrrrGrHrIrJrKrrfloatr~r1rLrnrMrErNrOrPs@rrrso/(/25/@C/MP/]b/mq/"%"3"3EDUDUZ_ZfZfrrc^\rSrSrSrS\S\S\S\S\S\S \S \S \4U4S jjr S \ RS\ R4Sjr SrU=r$)EfficientNetBlockiaA This corresponds to the expansion and depthwise convolution phase of each block in the original implementation. Args: config ([`EfficientNetConfig`]): Model configuration class. in_dim (`int`): Number of input channels. out_dim (`int`): Number of output channels. stride (`int`): Stride size to be used in convolution layers. expand_ratio (`int`): Expand ratio to set the output dimensions for the expansion and squeeze-excite layers. kernel_size (`int`): Kernel size for the depthwise convolution layer. drop_rate (`float`): Dropout rate to be used in the final phase of each block. id_skip (`bool`): Whether to apply dropout and sum the final hidden states with the input embeddings during the final phase of each block. Set to `True` for the first block of each stage. adjust_padding (`bool`): Whether to apply padding to only right and bottom side of the input kernel before the depthwise convolution operation, set to `True` for inputs with odd input sizes. rrar2r, expand_ratiorrrrrc f>[T U]5 XPlURS:gUlX%-n UR(a[ XXS9Ul[ UUR(aU OUUUU S9Ul[XXRS9Ul [UUR(aU OUUUUUS9Ul g)Nr)rrar2r,)rrar,rrr)rrarr)rrar2r,rr) r0r1rrr_ expansionrprvrsqueeze_exciter projection) r=rrar2r,rrrrrr expand_in_dimr>s rr1EfficientNetBlock.__init__s (''1, - ;;7mDN9$(KK=V#)  =];; 6$(KK=V  rrjrAcUnURS:waURU5nURU5nURU5nUR X!5nU$)Nr)rrrvrr)r=rjrs rrEEfficientNetBlock.forwardHsY"    ! NN=9M++M: ++M:  B r)rvrrrrrrrPs@rrrs4' "' '  '  '  ' ' ' ' ' R U%6%6 5<<  rrc v^\rSrSrSrS\4U4SjjrS S\RS\ \ S\ \ S\ 4S jjr S r U=r$) EfficientNetEncoderiUz Forward propagates the embeddings through each EfficientNet block. Args: config ([`EfficientNetConfig`]): Model configuration class. rc>^^[TT]5 UTlURTlU4Sjm[ UR 5n[ U4SjUR55nSn/n[U5Hn[XR U5n[XRU5nURUn URUn URUn [T"URU55Hcn U S:Hn U S:aSOU n U S:aUOUnXAR;nURU-U- n[!UUUU U U UU US9 nUR#U5 US- nMe M [$R&"U5Tl[$R*"W[US5SSSS 9Tl[$R."UR0UR2UR4S 9TlR:Tlg) Nc\>[[R"TRU-55$rC)rmathceildepth_coefficient)repeatsr=s r round_repeats3EfficientNetEncoder.__init__..round_repeatscs"tyy!7!7'!ABC Crc34># UH nT"U5v M g7frCr\).0nrs r /EfficientNetEncoder.__init__..hsL3Kaq))3Ksrr) rrar2r,rrrrrrircFrdrt)r0r1rrlenrUsumnum_block_repeatsrangerrVstrides kernel_sizes expand_ratiosdepthwise_paddingdrop_connect_raterappendr ModuleListblocksr4top_convr6 hidden_dimr7r8top_bnrr:top_activation)r=rnum_base_blocks num_blockscurr_block_numrirar2r,rrjrrrrblockrr>s` @rr1EfficientNetEncoder.__init__^s  !'!9!9 Df001L63K3KLL 'A"6+=+=a+@AF#F,?,?,BCG^^A&F --a0K!//2L=)A)A!)DEFq&!e$%Ev!/7O7O!O"44~E R )!!#! +!-'##1  e$!#'G(8mmF+  &vt4   nn**0E0EPVPjPj  %V%6%67rrjoutput_hidden_states return_dictrAcU(aU4OSnURHnU"U5nU(dMXA4- nM URU5nURU5nURU5nU(d[ SX455$[ UUS9$)Nc3.# UH ocMUv M g7frCr\)rvs rr.EfficientNetEncoder.forward..sX$Fq$Fs )last_hidden_staterj)rrrrtupler)r=rjrrall_hidden_statesrs rrEEfficientNetEncoder.forwards 1E],$[[E!-0M##!%55!!  m4  M2 ++M: X]$FXX X-++  r)rrrrrr)FT)rGrHrIrJrKrr1rLrnrr~rrErNrOrPs@rrrUs\58158t05&*  (( 'tn d^  (   rrcN\rSrSr%\\S'SrSr/rS\ R4Sjr Sr g) EfficientNetPreTrainedModelir efficientnetr@modulecV[U[R[R[R45(akUR R RSURRS9 URb%URR R5 ggg)zInitialize the weightsg)meanstdN) r rLinearr4r6weightdatanormal_rinitializer_ranger-zero_)r=rs r _init_weights)EfficientNetPreTrainedModel._init_weightssv fryy"))R^^D E E MM   & &CT[[5R5R & S{{&   &&(' Frr\N) rGrHrIrJr__annotations__base_model_prefixmain_input_name_no_split_modulesrModulerrNr\rrrrs) &$O)BII)rrc ^\rSrSrS\4U4Sjjr\S S\\RS\\ S\\ S\ \ \ 44Sjj5rS rU=r$) EfficientNetModelirc>[TU]U5 Xl[U5Ul[ U5UlURS:Xa%[R"URSS9Ul OMURS:Xa%[R"URSS9Ul O[SUR35eUR5 g)NrT) ceil_moderz2config.pooling must be one of ['mean', 'max'] got )r0r1rr$rrencoder pooling_typer AvgPool2drpooler MaxPool2d ValueErrorpooling post_initr<s rr1EfficientNetModel.__init__s   08*62    & (,,v'8'8DIDK  E ),,v'8'8DIDKQRXR`R`Qabc c rr@rrrAcUbUOURRnUbUOURRnUc [S5eUR U5nUR UUUS9nUSnUR U5nURURSS5nU(d Xg4USS-$[UUURS9$)Nz You have to specify pixel_valuesrrrrr)r pooler_outputrj) rruse_return_dictrrrrreshapeshaper rj)r=r@rrembedding_outputencoder_outputsr pooled_outputs rrEEfficientNetModel.forwards%9$D $++JjJj &1%> ? " " rrz EfficientNet Model with an image classification head on top (a linear layer on top of the pooled features), e.g. for ImageNet. ) custom_introc^\rSrSrU4Sjr\S S\\RS\\RS\\ S\\ S\ \ \ 44 Sjj5rS rU=r$) "EfficientNetForImageClassificationic>[TU]U5 URUlXl[ U5Ul[ R"URS9Ul URS:a+[ R"URUR5O[ R"5Ul UR5 g)Nrr)r0r1 num_labelsrrrrr dropout_raterrrIdentity classifierrr<s rr1+EfficientNetForImageClassification.__init__s   ++ -f5zzF$7$78 KO??]^K^"))F$5$5tGdfdododq rr@labelsrrrAcnUbUOURRnURXUS9nU(a UROUSnUR U5nUR U5nSnUbUR X'UR5nU(dU4USS-n UbU4U -$U $[UUURS9$)ab labels (`torch.LongTensor` of shape `(batch_size,)`, *optional*): Labels for computing the image classification/regression loss. Indices should be in `[0, ..., config.num_labels - 1]`. If `config.num_labels == 1` a regression loss is computed (Mean-Square loss), If `config.num_labels > 1` a classification loss is computed (Cross-Entropy). Nrrr)losslogitsrj) rrrrrr loss_functionr rj) r=r@rrroutputsr rroutputs rrE*EfficientNetForImageClassification.forward s&1%r(sc" " ! .,:   H % , C @U3:.@@(RYY2 ")) 44$$N$RYY$N"))BN NbW "))W t )/ ) ) 5 35 5 p / )D/  / d er