EiY* SSKrSSKJr SSKJrJrJr SSKrSSKJ r SSK J s J r SSKJ r SSKJr SSKJrJr SSKJr S S KJrJrJr S S KJr S S KJrJr S SKJ r J!r!J"r" /SQr#"SS\RH5r%"SS\RL5r'"SS\RP5r)"SS\RT5r+"SS\RX5r-"SS\R\5r/"SS\R`5r1"SS\Rd5r3"S S!\5r4\"S"S#9\"S$S%4S&9SS'S(S).S*\\\4\4S+\5S,\5S-\S.\34 S/jj55r6g)0N)partial)AnyOptionalUnion)Tensor) inception)Inception_V3_WeightsInceptionOutputs)ImageClassification)register_modelWeights WeightsEnum)_IMAGENET_CATEGORIES)_ovewrite_named_paramhandle_legacy_interface) _fuse_modules _replace_reluquantize_model)QuantizableInception3Inception_V3_QuantizedWeights inception_v3cf^\rSrSrS\S\SS4U4SjjrS\S\4SjrS S \\ SS4S jjr S r U=r $) QuantizableBasicConv2dargskwargsreturnNcZ>[TU]"U0UD6 [R"5UlgN)super__init__nnReLUreluselfrr __class__s f/home/mande/repo/quber/.venv/lib/python3.13/site-packages/torchvision/models/quantization/inception.pyr$QuantizableBasicConv2d.__init__s" $)&)GGI xclURU5nURU5nURU5nU$r"convbnr')r)r.s r+forwardQuantizableBasicConv2d.forwards. IIaL GGAJ IIaLr-is_qatc [U/SQUSS9 g)Nr0T)inplace)r)r)r5s r+ fuse_model!QuantizableBasicConv2d.fuse_model%sd2FDIr-)r'r") __name__ __module__ __qualname____firstlineno__rr$rr3rboolr8__static_attributes__ __classcell__r*s@r+rrsPcSTF J$J4JJr-rcJ^\rSrSrS\S\SS4U4SjjrS\S\4SjrS rU=r $) QuantizableInceptionA)rrr Ncx>[TU]"US[0UD6 [RR 5UlgN conv_blockr#r$rr% quantizedFloatFunctionalmyopr(s r+r$QuantizableInceptionA.__init__+/ $L+ALVLLL002 r-r.c\URU5nURRUS5$Nr_forwardrKcatr)r.outputss r+r3QuantizableInceptionA.forward/%--"yy}}Wa((r-rK r:r;r<r=rr$rr3r?r@rAs@r+rCrC)53c3S3T3))F))r-rCcJ^\rSrSrS\S\SS4U4SjjrS\S\4SjrS rU=r $) QuantizableInceptionB4rrr Ncx>[TU]"US[0UD6 [RR 5UlgrFrHr(s r+r$QuantizableInceptionB.__init__6rMr-r.c\URU5nURRUS5$rOrPrSs r+r3QuantizableInceptionB.forward:rVr-rWrXrAs@r+r[r[4rYr-r[cJ^\rSrSrS\S\SS4U4SjjrS\S\4SjrS rU=r $) QuantizableInceptionC?rrr Ncx>[TU]"US[0UD6 [RR 5UlgrFrHr(s r+r$QuantizableInceptionC.__init__ArMr-r.c\URU5nURRUS5$rOrPrSs r+r3QuantizableInceptionC.forwardErVr-rWrXrAs@r+rbrb?rYr-rbcJ^\rSrSrS\S\SS4U4SjjrS\S\4SjrS rU=r $) QuantizableInceptionDJrrr Ncx>[TU]"US[0UD6 [RR 5UlgrFrHr(s r+r$QuantizableInceptionD.__init__LrMr-r.c\URU5nURRUS5$rOrPrSs r+r3QuantizableInceptionD.forwardPrVr-rWrXrAs@r+ririJrYr-ricb^\rSrSrS\S\SS4U4SjjrS\S\\4SjrS\S\4S jr S r U=r $) QuantizableInceptionEUrrr Nc>[TU]"US[0UD6 [RR 5Ul[RR 5Ul[RR 5UlgrF) r#r$rr%rIrJmyop1myop2myop3r(s r+r$QuantizableInceptionE.__init__WsW $L+ALVL\\113 \\113 \\113 r-r.cURU5nURU5nURU5URU5/nURR US5nUR U5nURU5nURU5URU5/nURR US5n[R"USSSS9nURU5nX#XE/nU$)Nrr ) kernel_sizestridepadding) branch1x1 branch3x3_1 branch3x3_2a branch3x3_2brsrRbranch3x3dbl_1branch3x3dbl_2branch3x3dbl_3abranch3x3dbl_3brtF avg_pool2d branch_pool)r)r.r{ branch3x3 branch3x3dblrrTs r+rQQuantizableInceptionE._forward]sNN1% $$Q' &&y143D3DY3OP JJNN9a0 **1- **<8   .   . zz~~lA6 ll1!AqI &&{3 Cr-c\URU5nURRUS5$rO)rQrurRrSs r+r3QuantizableInceptionE.forwardrs%--"zz~~gq))r-)rsrtru) r:r;r<r=rr$rlistrQr3r?r@rAs@r+rprpUsL4c4S4T4 &T&\***F**r-rpc8^\rSrSrS\S\SS4U4SjjrSrU=r$)QuantizableInceptionAuxwrrr Nc2>[TU]"US[0UD6 grF)r#r$rr(s r+r$ QuantizableInceptionAux.__init__ys $L+ALVLr-)r:r;r<r=rr$r?r@rAs@r+rrws'McMSMTMMr-rcf^\rSrSrS\S\SS4U4SjjrS\S\4SjrS S \ \ SS4S jjr S r U=r $) r}rrr Nc $>[TU]"US[[[[ [ [[/0UD6 [RRR5Ul [RRR5Ulg)Ninception_blocks)r#r$rrCr[rbrirprtorchao quantization QuantStubquant DeQuantStubdequantr(s r+r$QuantizableInception3.__init__~st   '%%%%%'  XX**446 xx,,88: r-r.cURU5nURU5nURU5upURU5nUR=(a UR n[ RR5(a(U(d[R"S5 [X5$URX5$)NzIScripted QuantizableInception3 always returns QuantizableInception3 Tuple) _transform_inputrrQrtraining aux_logitsrjit is_scriptingwarningswarnr eager_outputs)r)r.aux aux_defineds r+r3QuantizableInception3.forwards  ! !! $ JJqMq! LLOmm7 99 ! ! # # ij#A+ +%%a- -r-r5c|UR5H(n[U5[LdMURU5 M* g)zFuse conv/bn/relu modules in inception model Fuse conv+bn+relu/ conv+relu/conv+bn modules to prepare for quantization. Model is modified in place. Note that this operation does not change numerics and the model after modification is in floating point N)modulestyperr8)r)r5ms r+r8 QuantizableInception3.fuse_models-AAw00 V$ r-)rrr")r:r;r<r=rr$rr r3rr>r8r?r@rAs@r+rr}sM;c;S;T;" . .$4 . %$ %4 % %r-rch\rSrSr\"S\"\SSS9SS\SS \RS S S S .0SSSS. S9r \ r Sr g)rzUhttps://download.pytorch.org/models/quantized/inception_v3_google_fbgemm-a2837893.pthi+iV) crop_size resize_sizeir)Krfbgemmzdhttps://github.com/pytorch/vision/tree/main/references/classification#post-training-quantized-modelsz ImageNet-1Kg%CKS@g-VW@)zacc@1zacc@5g'1@gL7A`%7@z These weights were produced by doing Post Training Quantization (eager mode) on top of the unquantized weights listed below. ) num_paramsmin_size categoriesbackendrecipe unquantized_metrics_ops _file_size_docs)url transformsmetarN) r:r;r<r=rrr rr IMAGENET1K_V1IMAGENET1K_FBGEMM_V1DEFAULTr?rr-r+rrsa" c.#3O" .|/==##    0#Gr-rquantized_inception_v3)name pretrainedcpURSS5(a[R$[R$)NquantizeF)getrrr r)rs r+rs1zz*e,, * > > 4%33 4r-)weightsTF)rprogressrrrrrr c vU(a[O[RU5nURSS5nUbmSU;a [ USS5 [ USS5 [ US[ UR S55 SUR ;a[ USUR S5 URSS 5n[S 0UD6n[U5 U(a [Xe5 UbWU(aU(dSUl SUl URURUSS 95 U(dU(dSUl SUl U$) a6Inception v3 model architecture from `Rethinking the Inception Architecture for Computer Vision `__. .. note:: **Important**: In contrast to the other models the inception_v3 expects tensors with a size of N x 3 x 299 x 299, so ensure your images are sized accordingly. .. note:: Note that ``quantize = True`` returns a quantized model with 8 bit weights. Quantized models only support inference and run on CPUs. GPU inference is not yet supported. Args: weights (:class:`~torchvision.models.quantization.Inception_V3_QuantizedWeights` or :class:`~torchvision.models.Inception_V3_Weights`, optional): The pretrained weights for the model. See :class:`~torchvision.models.quantization.Inception_V3_QuantizedWeights` 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. quantize (bool, optional): If True, return a quantized version of the model. Default is False. **kwargs: parameters passed to the ``torchvision.models.quantization.QuantizableInception3`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.quantization.Inception_V3_QuantizedWeights :members: .. autoclass:: torchvision.models.Inception_V3_Weights :members: :noindex: rFNtransform_inputT num_classesrrr)r check_hashr)rr verifyrrlenrpoprrrr AuxLogitsload_state_dictget_state_dict)rrrroriginal_aux_logitsrmodels r+rrsh19,>RZZ[bcG **\59 F * !&*;T BflD9fmSl9S5TU  $ !&)W\\)5L MjjH-G ! +F +E%u& /$E "EO g44hSW4XY 3$E "EO Lr-)7r functoolsrtypingrrrrtorch.nnr%torch.nn.functional functionalrrtorchvision.modelsrinception_moduletorchvision.models.inceptionr r transforms._presetsr _apirrr_metar_utilsrrutilsrrr__all__ BasicConv2dr InceptionArC InceptionBr[ InceptionCrb InceptionDri InceptionErp InceptionAuxr Inception3rrr>rrr-r+rsu'' <O677(C??  J-99 J),77)),77)),77)),77)*,77*DM.;;M )%,77)%X#K#8-.   UY C e9;OOP QCC C  C  C /Cr-