kh #ddlmZddlmZmZddlmZddlmZddl m Z m Z m Z ddl mZdd lmZmZmZdd lmZmZmZmZmZd d lmZgd ZGddeZGddej8ZedddZGdde ZGdde Z dede!dee"defdZ#e edejHfdejJf d!d"d!d!ejJd#d$eed%e"dee!d&ee"d'eed(edefd)Z&e ede jHfdejJf d!d"d!d!ejJd#d$ee d%e"dee!d&ee"d'eed(edefd*Z'y!)+)partial)AnyOptional)nn)SemanticSegmentation)register_modelWeights WeightsEnum)_VOC_CATEGORIES)_ovewrite_value_paramhandle_legacy_interfaceIntermediateLayerGetter)ResNet resnet101ResNet101_Weightsresnet50ResNet50_Weights)_SimpleSegmentationModel)FCNFCN_ResNet50_WeightsFCN_ResNet101_Weights fcn_resnet50 fcn_resnet101ceZdZdZy)ra Implements FCN model from `"Fully Convolutional Networks for Semantic Segmentation" `_. 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 "out" for the last feature map used, and "aux" if an auxiliary classifier is used. classifier (nn.Module): module that takes the "out" element returned from the backbone and returns a dense prediction. aux_classifier (nn.Module, optional): auxiliary classifier used during training N)__name__ __module__ __qualname____doc___/var/www/teggl/fontify/venv/lib/python3.12/site-packages/torchvision/models/segmentation/fcn.pyrrs   r#rc,eZdZdededdffd ZxZS)FCNHead in_channelschannelsreturnNc |dz}tj||dddtj|tjtjdtj||dg}t ||y)NrrF)paddingbiasg?)rConv2d BatchNorm2dReLUDropoutsuper__init__)selfr'r(inter_channelslayers __class__s r$r3zFCNHead.__init__%sg$) IIk>1ae L NN> * GGI JJsO IInh 2   &!r#)rrr intr3 __classcell__)r7s@r$r&r&$s" "C "3 "4 " "r#r&)rrz These weights were trained on a subset of COCO, using only the 20 categories that are present in the Pascal VOC dataset. ) categoriesmin_size_docsc TeZdZedeediedddddd id d d  ZeZy)rzBhttps://download.pytorch.org/models/fcn_resnet50_coco-1167a1af.pth resize_sizeijzPhttps://github.com/pytorch/vision/tree/main/references/segmentation#fcn_resnet50COCO-val2017-VOC-labelsg@N@gV@miou pixel_accgmc@g?5^I`@ num_paramsrecipe_metrics_ops _file_sizeurl transformsmetaN rrr r rr _COMMON_METACOCO_WITH_VOC_LABELS_V1DEFAULTr"r#r$rr<sU% P/SA    "h) !%, !  "&Gr#rc TeZdZedeediedddddd id d d  ZeZy)rzChttps://download.pytorch.org/models/fcn_resnet101_coco-7ecb50ca.pthr>r?ij<zWhttps://github.com/pytorch/vision/tree/main/references/segmentation#deeplabv3_resnet101rAgO@gV@rBgV-m@gni@rErKNrOr"r#r$rrQsU% Q/SA    "o) !%, !  "&Gr#rbackbone num_classesauxr)cddi}|rd|d<t||}|r td|nd}td|}t|||S)Nlayer4outrVlayer3) return_layersii)rr&r)rTrUrVr[aux_classifier classifiers r$ _fcn_resnetr^fsR u%M "' h&x}MH36WT;/DN{+J x^ 44r# pretrainedpretrained_backbone)weightsweights_backboneNT)raprogressrUaux_lossrbrarcrdrbkwargsc Ltj|}tj|}|3d}td|t |j d}td|d}n|d}t |gd}t|||}|"|j|j|d |S) a\Fully-Convolutional Network model with a ResNet-50 backbone from the `Fully Convolutional Networks for Semantic Segmentation `_ paper. .. betastatus:: segmentation module Args: weights (:class:`~torchvision.models.segmentation.FCN_ResNet50_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.segmentation.FCN_ResNet50_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.ResNet50_Weights`, optional): The pretrained weights for the backbone. **kwargs: parameters passed to the ``torchvision.models.segmentation.fcn.FCN`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.segmentation.FCN_ResNet50_Weights :members: NrUr:rdTFTTrareplace_stride_with_dilationrc check_hash) rverifyrrlenrNrr^load_state_dictget_state_dictrarcrUrdrbrerTmodels r$rrusP#))'2G'../?@+M;GLLYeLfHgh (XtD    0ObcH +x 8E g44hSW4XY Lr#c Ltj|}tj|}|3d}td|t |j d}td|d}n|d}t |gd}t|||}|"|j|j|d |S) aaFully-Convolutional Network model with a ResNet-101 backbone from the `Fully Convolutional Networks for Semantic Segmentation `_ paper. .. betastatus:: segmentation module Args: weights (:class:`~torchvision.models.segmentation.FCN_ResNet101_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.segmentation.FCN_ResNet101_Weights` below for more details, and possible values. By default, no pre-trained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. num_classes (int, optional): number of output classes of the model (including the background). aux_loss (bool, optional): If True, it uses an auxiliary loss. weights_backbone (:class:`~torchvision.models.ResNet101_Weights`, optional): The pretrained weights for the backbone. **kwargs: parameters passed to the ``torchvision.models.segmentation.fcn.FCN`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.segmentation.FCN_ResNet101_Weights :members: NrUr:rdTrgrhrirk) rrmrrrnrNrr^rorprqs r$rrsP$**73G(//0@A+M;GLLYeLfHgh (XtD   !1PcdH +x 8E g44hSW4XY Lr#)( functoolsrtypingrrtorchrtransforms._presetsr_apir r r _metar _utilsrrrresnetrrrrrr__all__r Sequentialr&rPrrr8boolr^rQ IMAGENET1K_V1rrr"r#r$rs  777#\\UU, d " & "bmm "" &;&*&K&* 5 5 5 $ 5 5 /GG H+-=-K-KL /3!%#3C3Q3Q 3 * +33# 3 tn 3 /0 33 3  3l 0HH I+->-L-LM 04!%#4E4S4S 3 + ,33# 3 tn 3 01 33 3  3r#