+ j ^RIHt^RIHt^RIHtHt^RIHtH t ^RI H t ^RI Ht^RIHt^RIHtHtHt^R IHt^R IHtHtHt^R IHtHtHt.ROt!RR ]P@4t!!RR]P@4t"RRlt#!RR ]4t$]!4]!R]$PJ3R]PL3R7RRRRRRR]PL/RRll44t'R#) ) OrderedDict)partial)AnyOptional)nnTensor) functional)SemanticSegmentation)_log_api_usage_once)register_modelWeights WeightsEnum)_VOC_CATEGORIES)_ovewrite_value_paramhandle_legacy_interfaceIntermediateLayerGetter)mobilenet_v3_largeMobileNet_V3_Large_Weights MobileNetV3LRASPP!LRASPP_MobileNet_V3_Large_WeightscXaa]tRt^toRtRV3RlV3RllltV3RlRltRtVtV;t #)ra Implements a Lite R-ASPP Network for semantic segmentation from `"Searching for MobileNetV3" `_. Args: backbone (nn.Module): the network used to compute the features for the model. The backbone should return an OrderedDict[Tensor], with the key being "high" for the high level feature map and "low" for the low level feature map. low_channels (int): the number of channels of the low level features. high_channels (int): the number of channels of the high level features. num_classes (int, optional): number of output classes of the model (including the background). inter_channels (int, optional): the number of channels for intermediate computations. c P<V^8dQhRS[PRS[RS[RS[RS[RR/#)backbone low_channels high_channels num_classesinter_channelsreturnN)rModuleint)format __classdict__s"g/home/luca/screen-checking/.venv/lib/python3.14/site-packages/torchvision/models/segmentation/lraspp.py __annotate__LRASPP.__annotate__#sI__ _14_EH_WZ_lo_ _ch<\SV`4\V4Wn\ W#WE4VnR#)N)super__init__r r LRASPPHead classifier)selfrrrrr __class__s&&&&&&r%r+LRASPP.__init__#s+ D! $\+^r(c<<V^8dQhRS[RS[S[S[3,/#rinputr )rdictstr)r#r$s"r%r&r'+s$VS&[(9r(cVPV4pVPV4p\P!W1PRRRRR7p\ 4pW4R&V#)rNbilinearFsizemode align_cornersout)rr-F interpolateshaper)r.r3featuresr<results&& r%forwardLRASPP.forward+sO=='ooh'mmCkk"#&6ZW\]u  r()rr-)) __name__ __module__ __qualname____firstlineno____doc__r+rC__static_attributes____classdictcell__ __classcell__r/r$s@@r%rrs% __r(cPaa]tRt^6toV3RlV3RlltV3RlRltRtVtV;t#)r,c 6<V^8dQhRS[RS[RS[RS[RR/#)rrrrrr N)r")r#r$s"r%r&LRASPPHead.__annotate__7s9 I IS I I3 I`c Ihl Ir(c <\SV`4\P!\P!W$^RR7\P !V4\P !RR74Vn\P!\P!^4\P!W$^RR7\P!44Vn \P!W^4Vn \P!WC^4Vn R#)F)biasT)inplaceN) r*r+r SequentialConv2d BatchNorm2dReLUcbrAdaptiveAvgPool2dSigmoidscalelow_classifierhigh_classifier)r.rrrrr/s&&&&&r%r+LRASPPHead.__init__7s == IImQU C NN> * GGD !  ]]   # IImQU C JJL  !ii 1E!yyaHr(c<<V^8dQhRS[S[S[3,RS[/#r2)r4r5r)r#r$s"r%r&rQFs( B BT#v+. B6 Br(cVR,pVR,pVPV4pVPV4pWE,p\P!WBPRRRRR7pVP V4VP V4,#)lowhighNr7Fr8r=)rZr]r>r?r@r^r_)r.r3rcrdxss&& r%rCLRASPPHead.forwardFstElV} HHTN JJt  E MM!))BC.zQV W""3'$*>*>q*AAAr()rZr_r^r]) rFrGrHrIr+rCrKrLrMrNs@@r%r,r,6s# I I B B Br(r,c<V^8dQhR\R\R\/#)rrrr )rr"r)r#s"r%r&r&Rs& F F+ FC FF Fr(c VPp^.\V4UUu.uFwr#\VRR4'gKVNK upp,\V4^, .,pVR,pVR,pW,PpW,Pp\ V\ V4R\ V4R/R7p\WW4#uuppi)r_is_cnFrcrd) return_layers)rA enumerategetattrlen out_channelsrr5r) rrib stage_indiceslow_poshigh_posrrs && r%_lraspp_mobilenetv3rwRs  HC8)<\)<8UZ@[11)<\\`cdl`mpq`q_rrMBGR H$11L&33M&xG eUXYaUbdj?klH (- EE]s CCcl]tRt^`t]!R]!]RR7RRR]RRRR R R R R RR//RRRRRR/R7t]t Rt R#)rzJhttps://download.pytorch.org/models/lraspp_mobilenet_v3_large-d234d4ea.pthi) resize_size num_paramsi"(1 categoriesmin_sizerecipez]https://github.com/pytorch/vision/tree/main/references/segmentation#lraspp_mobilenet_v3_large_metricszCOCO-val2017-VOC-labelsmioug33333L@ pixel_accgV@_opsg㥛 @ _file_sizeg{G(@_docsz These weights were trained on a subset of COCO, using only the 20 categories that are present in the Pascal VOC dataset. )url transformsmetaN)rSrS) rFrGrHrIr rr rCOCO_WITH_VOC_LABELS_V1DEFAULTrKrr(r%rr`sj% X/SA ' /  u )D, E %  ,&Gr( pretrainedpretrained_backbone)weightsweights_backbonerNprogressTrrc V^8dQhR\\,R\R\\,R\\,R\ R\ /#)rrrrrkwargsr )rrboolr"rrr)r#s"r%r&r&sV33 7 833# 3 9: 3  3 3r(c VPRR4'd \R4h\PV4p\P!V4pVe+Rp\ RV\ VPR,44pMVf^p\VRR7p\WR4pVe#VPVPVRR 74V#) a(Constructs a Lite R-ASPP Network model with a MobileNetV3-Large backbone from `Searching for MobileNetV3 `_ paper. .. betastatus:: segmentation module Args: weights (:class:`~torchvision.models.segmentation.LRASPP_MobileNet_V3_Large_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.segmentation.LRASPP_MobileNet_V3_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. num_classes (int, optional): number of output classes of the model (including the background). aux_loss (bool, optional): If True, it uses an auxiliary loss. weights_backbone (:class:`~torchvision.models.MobileNet_V3_Large_Weights`, optional): The pretrained weights for the backbone. **kwargs: parameters passed to the ``torchvision.models.segmentation.LRASPP`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.segmentation.LRASPP_MobileNet_V3_Large_Weights :members: aux_lossFz&This model does not use auxiliary lossNrr{T)rdilated)r check_hash) popNotImplementedErrorrverifyrrrprrrwload_state_dictget_state_dict)rrrrrrmodels$$$$, r%lraspp_mobilenet_v3_largerzsLzz*e$$!"JKK/66w?G1889IJ+M;GLLYeLfHgh   !*:DIH  6E g44hSW4XY Lr()rrr)( collectionsr functoolsrtypingrrtorchrrtorch.nnr r>transforms._presetsr utilsr _apir r r_metar_utilsrrr mobilenetv3rrr__all__r!rr,rwrr IMAGENET1K_V1rrr(r%rs# $7(77#\\UU W RYY FBB8 F& &4 <TT U+-G-U-UV3;?33"& 3 >X=e=e 3  3r(