U hA @sddlZddlmZddlmZddlmZmZmZm Z ddl Z ddl m Z ddl m mZddlmmZddl mZddlmZddlmZd d lmZmZmZd d lmZd d lmZm Z d ddddddddg Z!Gddde j"Z#Gddde j$Z%Gddde j&Z'Gdd d e j"Z(e j"ee)ddddZ*e+e e+e+e+e+fe+eee)ee(d d!d"Z,d#ed$d%d&Z-Gd'ddeZ.Gd(ddeZ/Gd)ddeZ0Gd*ddeZ1ee d+e.j2fd,dd-d.ee.e)ee(d/d0dZ3ee d+e/j2fd,dd-d.ee/e)ee(d/d1dZ4ee d+e0j2fd,dd-d.ee0e)ee(d/d2dZ5ee d+e1j2fd,dd-d.ee1e)ee(d/d3dZ6dS)4N) OrderedDict)partial)AnyListOptionalTuple)Tensor)ImageClassification)_log_api_usage_once)register_modelWeights WeightsEnum)_IMAGENET_CATEGORIES)_ovewrite_named_paramhandle_legacy_interfaceDenseNetDenseNet121_WeightsDenseNet161_WeightsDenseNet169_WeightsDenseNet201_Weights densenet121 densenet161 densenet169 densenet201cseZdZdeeeeeddfdd ZeeedddZ eeed d d Z e j j eeed d d Ze j jeeed ddZe j jeed ddZeed ddZZS) _DenseLayerFN)num_input_features growth_ratebn_size drop_ratememory_efficientreturncstt||_tjdd|_tj|||dddd|_t|||_ tjdd|_ tj|||ddddd|_ t ||_ ||_dS)NTZinplacer F kernel_sizestridebiasr%r&paddingr')super__init__nn BatchNorm2dnorm1ReLUrelu1Conv2dconv1norm2relu2conv2floatr r!)selfrrrr r! __class__M/var/www/html/venv/lib/python3.8/site-packages/torchvision/models/densenet.pyr, s   z_DenseLayer.__init__)inputsr"cCs&t|d}||||}|SNr )torchcatr3r1r/)r8r=Zconcated_featuresbottleneck_outputr;r;r< bn_function/s z_DenseLayer.bn_function)inputr"cCs|D]}|jrdSqdS)NTF)Z requires_grad)r8rCZtensorr;r;r<any_requires_grad5sz_DenseLayer.any_requires_gradcs"fdd}tj|f|ddiS)Ncs |SN)rB)r=r8r;r<closure=sz7_DenseLayer.call_checkpoint_bottleneck..closureZ use_reentrantF)cp checkpoint)r8rCrGr;rFr<call_checkpoint_bottleneck;s z&_DenseLayer.call_checkpoint_bottleneckcCsdSrEr;r8rCr;r;r<forwardBsz_DenseLayer.forwardcCsdSrEr;rKr;r;r<rLFscCst|tr|g}n|}|jrD||rDtjr8td||}n | |}| | | |}|j dkrtj||j |jd}|S)Nz%Memory Efficient not supported in JITr)ptraining) isinstancerr!rDr?jitZ is_scripting ExceptionrJrBr6r5r4r FZdropoutrN)r8rCZ prev_featuresrA new_featuresr;r;r<rLLs     )F)__name__ __module__ __qualname__intr7boolr,rrrBrDr?rPZunusedrJZ_overload_methodrL __classcell__r;r;r9r<rs$rc sDeZdZdZd eeeeeeddfdd Zeeddd Z Z S) _DenseBlockr FN) num_layersrrrr r!r"c sJtt|D]2}t|||||||d}|d|d|qdS)N)rrr r!z denselayer%dr )r+r,ranger add_module) r8r[rrrr r!ilayerr9r;r<r,cs   z_DenseBlock.__init__) init_featuresr"cCs6|g}|D]\}}||}||qt|dSr>)itemsappendr?r@)r8r`featuresnamer_rSr;r;r<rLws  z_DenseBlock.forward)F) rTrUrV_versionrWr7rXr,rrLrYr;r;r9r<rZ`s rZcs&eZdZeeddfdd ZZS) _TransitionN)rnum_output_featuresr"csNtt||_tjdd|_tj||dddd|_tj ddd|_ dS)NTr#r Fr$r )r%r&) r+r,r-r.Znormr0relur2convZ AvgPool2dpool)r8rrgr9r;r<r,s   z_Transition.__init__)rTrUrVrWr,rYr;r;r9r<rfsrfc sReZdZdZdeeeeeefeeeeed d fd d Ze e d ddZ Z S)raKDensenet-BC model class, based on `"Densely Connected Convolutional Networks" `_. Args: growth_rate (int) - how many filters to add each layer (`k` in paper) block_config (list of 4 ints) - how many layers in each pooling block num_init_features (int) - the number of filters to learn in the first convolution layer bn_size (int) - multiplicative factor for number of bottle neck layers (i.e. bn_size * k features in the bottleneck layer) drop_rate (float) - dropout rate after each dense layer num_classes (int) - number of classification classes memory_efficient (bool) - If True, uses checkpointing. Much more memory efficient, but slower. Default: *False*. See `"paper" `_.  @rFN)r block_confignum_init_featuresrr num_classesr!r"c stt|ttdtjd|dddddfdt|fdtjd d fd tj ddd d fg|_ |}t |D]|\} } t | |||||d} |j d| d | || |}| t|d krrt||dd} |j d| d | |d}qr|j dt|t|||_|D]r} t| tjr||}tj|dd}t|d}t|d}||}|S)NTr#)r r r )rcrRrhZadaptive_avg_pool2dr?flattenr{)r8rrcoutr;r;r<rLs     zDenseNet.forward)rkrlrqrrrrsF) rTrUrV__doc__rWrr7rXr,rrLrYr;r;r9r<rs&<)modelweightsprogressr"cCsltd}|j|dd}t|D]8}||}|r$|d|d}||||<||=q$||dS)Nz]^(.*denselayer\d+\.(?:norm|relu|conv))\.((?:[12])\.(?:weight|bias|running_mean|running_var))$T)rZ check_hashr r )recompileZget_state_dictlistkeysmatchgroupZload_state_dict)rrrpatternZ state_dictkeyresZnew_keyr;r;r<_load_state_dicts  r)rrtrurrkwargsr"cKsH|dk rt|dt|jdt|||f|}|dk rDt|||d|S)Nrv categories)rrr)rrzmetarr)rrtrurrrrr;r;r< _densenets r)rz*https://github.com/pytorch/vision/pull/116z'These weights are ported from LuaTorch.)Zmin_sizerZrecipeZ_docsc @s>eZdZedeeddedddddid d d d ZeZd S)rz@Z num_paramsZ_metricsZ_ops _file_sizeurlZ transformsrN rTrUrVrrr _COMMON_META IMAGENET1K_V1DEFAULTr;r;r;r<r s c @s>eZdZedeeddedddddid d d d ZeZd S)rzeZdZedeeddedddddid d d d ZeZd S)rzeZdZedeeddedddddid d d d ZeZd S)rz`_. Args: weights (:class:`~torchvision.models.DenseNet121_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.DenseNet121_Weights` below for more details, and possible values. By default, no pre-trained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to the ``torchvision.models.densenet.DenseNet`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.DenseNet121_Weights :members: rkrlrq)rverifyrrrrr;r;r<r[s cKst|}tddd||f|S)a{Densenet-161 model from `Densely Connected Convolutional Networks `_. Args: weights (:class:`~torchvision.models.DenseNet161_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.DenseNet161_Weights` below for more details, and possible values. By default, no pre-trained weights are used. progress (bool, optional): If True, displays a progress bar of the download to stderr. Default is True. **kwargs: parameters passed to the ``torchvision.models.densenet.DenseNet`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.DenseNet161_Weights :members: 0)rmrn$ro`)rrrrr;r;r<rus cKst|}tddd||f|S)a{Densenet-169 model from `Densely Connected Convolutional Networks `_. Args: weights (:class:`~torchvision.models.DenseNet169_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.DenseNet169_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.densenet.DenseNet`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.DenseNet169_Weights :members: rk)rmrnrkrkrq)rrrrr;r;r<rs cKst|}tddd||f|S)a{Densenet-201 model from `Densely Connected Convolutional Networks `_. Args: weights (:class:`~torchvision.models.DenseNet201_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.DenseNet201_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.densenet.DenseNet`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.DenseNet201_Weights :members: rk)rmrnrrkrq)rrrrr;r;r<rs )7r collectionsr functoolsrtypingrrrrr?Ztorch.nnr-Ztorch.nn.functionalZ functionalrRZtorch.utils.checkpointutilsrIrHrZtransforms._presetsr r Z_apir rr_metarZ_utilsrr__all__ModulerZ ModuleDictrZrxrfrrXrrWrrrrrrrrrrrr;r;r;r<sr        A U $$$