a PSic!4 @sRddlZddlmZddlmZddlmZmZmZm Z m Z m Z ddl Z ddl m Z mZddlmZddlmZdd lmZd d lmZd d lmZmZd d lmZd dlmZmZd dlm Z m!Z!ddlmZ"ddl#m$Z$ddl%m&Z&ddl'm(Z(m)Z)ddgZ*e+e+e+ede j,fe j-dddZ.e+e+ede j,fe j-dddZ/e j,ddd Z0Gd!d"d"e j,Z1Gd#d$d$e)Z2Gd%d&d&e)Z3Gd'd(d(e j,Z4e ej5ej6fe+ede j,fd)d*d+Z7Gd,ddeZ8ed-e8j9fd.e j:fd/dd0de j:ddd1e e8e;e e+e e e e+e ede j,fee(d2d3dZiZ?dS)6N) OrderedDict)partial)AnyCallableDictListOptionalUnion)nnTensor)Conv2dNormActivation)ObjectDetection)_log_api_usage_once) mobilenet) WeightsEnumWeights)_COCO_CATEGORIES)handle_legacy_interface_ovewrite_value_param)MobileNet_V3_Large_Weightsmobilenet_v3_large)_utils)DefaultBoxGenerator)_validate_trainable_layers)SSDSSDScoringHead%SSDLite320_MobileNet_V3_Large_Weightsssdlite320_mobilenet_v3_large.) in_channels out_channels kernel_size norm_layerreturnc Cs(tt|||||tjdt||dS)N)r#groupsr$activation_layerr)r Sequentialr ReLU6Conv2d)r!r"r#r$r+`/var/www/html/django/DPS/env/lib/python3.9/site-packages/torchvision/models/detection/ssdlite.py_prediction_blocks r-)r!r"r$r%c CsJtj}|d}tt||d||dt||dd|||dt||d||dS)Nrr)r#r$r'r )r#strider&r$r')r r)r(r )r!r"r$ activationZintermediate_channelsr+r+r, _extra_block0s&   r0)convcCsP|D]B}t|tjrtjjj|jddd|jdurtjj |jdqdS)NgQ?)meanstd) modules isinstancer r*torchinitnormal_weightbias constant_)r1layerr+r+r, _normal_initIs    r>csVeZdZeeeeeedejfdfdd Zee e e e fdddZ Z S) SSDLiteHead.r! num_anchors num_classesr$cs,tt|||||_t||||_dSN)super__init__SSDLiteClassificationHeadclassification_headSSDLiteRegressionHeadregression_head)selfr!rArBr$ __class__r+r,rERs zSSDLiteHead.__init__xr%cCs||||dS)N)bbox_regression cls_logits)rIrG)rJrNr+r+r,forwardYszSSDLiteHead.forward)__name__ __module__ __qualname__rintrr ModulerEr rstrrQ __classcell__r+r+rKr,r?Qsr?cs:eZdZeeeeeedejfdfdd ZZ S)rF.r@csNt}t||D] \}}|t|||d|qt|t||dS)Nr r ModuleListzipappendr-r>rDrE)rJr!rArBr$rPchannelsanchorsrKr+r,rEas z"SSDLiteClassificationHead.__init__ rRrSrTrrUrr rVrErXr+r+rKr,rF`srFcs8eZdZeeeeedejfdfdd ZZ S)rH.)r!rAr$csNt}t||D] \}}|t|d|d|qt|t|ddS)Nr rY)rJr!rAr$bbox_regr]r^rKr+r,rEls zSSDLiteRegressionHead.__init__r_r+r+rKr,rHksrHcsPeZdZd ejeedejfeedfdd Ze e e e fddd Z Z S) SSDLiteFeatureExtractorMobileNet?.)backbonec4_posr$ width_mult min_depthc stt|||jr$tdttjg|d|||jdRtj||jddg||ddR|_fdd}t t |dj |d|t |d|d|t |d|d|t |d|d |g}t |||_ dS) Nz0backbone[c4_pos].use_res_connect should be Falserrcstt|SrC)maxrU)drhrgr+r,z;SSDLiteFeatureExtractorMobileNet.__init__..i)rDrEruse_res_connect ValueErrorr r(blockfeaturesrZr0r"r>extra)rJrerfr$rgrhZ get_depthrurKrkr,rEus$  $*z)SSDLiteFeatureExtractorMobileNet.__init__rMcCsVg}|jD]}||}||q |jD]}||}||q(tddt|DS)NcSsg|]\}}t||fqSr+)rW).0ivr+r+r, rmz.)rtr\rur enumerate)rJrNoutputrsr+r+r,rQs    z(SSDLiteFeatureExtractorMobileNet.forward)rcrd)rRrSrTr rVrUrfloatrEr rrWrQrXr+r+rKr,rbts !rb)retrainable_layersr$cCs|j}dgddt|Dt|dg}t|}d|krH|ksRntd|dkrbt|n |||}|d|D]}|D]}|dqqzt||d|S)NrcSs g|]\}}t|ddr|qS)_is_cnF)getattr)rvrwbr+r+r,ryrmz(_mobilenet_extractor..rzYtrainable_layers should be in the range [0, {num_stages}], instead got {trainable_layers}F)rtrzlenrr parametersrequires_grad_rb)rer}r$ stage_indices num_stages freeze_beforer parameterr+r+r,_mobilenet_extractors& rc @s4eZdZedededddddiidd d ZeZd S) rzShttps://download.pytorch.org/models/ssdlite320_mobilenet_v3_large_coco-a79551df.pthi}4)rrz]https://github.com/pytorch/vision/tree/main/references/detection#ssdlite320-mobilenetv3-largez COCO-val2017box_mapgL5@zSThese weights were produced by following a similar training recipe as on the paper.) num_params categoriesmin_sizerecipe_metrics_docs)url transformsmetaN)rRrSrTrrrCOCO_V1DEFAULTr+r+r+r,rs pretrainedpretrained_backbone)weightsweights_backboneT)rprogressrBrtrainable_backbone_layersr$)rrrBrrr$kwargsr%cKst|}t|}d|vr&td|durHd}t|t|jd}n |durTd}t|dupd|du|dd}|du}|durt t j ddd }t f||||d |}|durt |t|||}d } td d tdDddd} t|| } | } t| t| jkr(tdt| dt| jddddgdgdd} i| |}t|| | |fdt| | ||i|}|dur||j|d|S)aSSDlite model architecture with input size 320x320 and a MobileNetV3 Large backbone, as described at `Searching for MobileNetV3 `__ and `MobileNetV2: Inverted Residuals and Linear Bottlenecks `__. .. betastatus:: detection module See :func:`~torchvision.models.detection.ssd300_vgg16` for more details. Example: >>> model = torchvision.models.detection.ssdlite320_mobilenet_v3_large(weights=SSDLite320_MobileNet_V3_Large_Weights.DEFAULT) >>> model.eval() >>> x = [torch.rand(3, 320, 320), torch.rand(3, 500, 400)] >>> predictions = model(x) Args: weights (:class:`~torchvision.models.detection.SSDLite320_MobileNet_V3_Large_Weights`, optional): The pretrained weights to use. See :class:`~torchvision.models.detection.SSDLite320_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). weights_backbone (:class:`~torchvision.models.MobileNet_V3_Large_Weights`, optional): The pretrained weights for the backbone. trainable_backbone_layers (int, optional): number of trainable (not frozen) layers starting from final block. Valid values are between 0 and 6, with 6 meaning all backbone layers are trainable. If ``None`` is passed (the default) this value is set to 6. norm_layer (callable, optional): Module specifying the normalization layer to use. **kwargs: parameters passed to the ``torchvision.models.detection.ssd.SSD`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.detection.SSDLite320_MobileNet_V3_Large_Weights :members: sizez?The size of the model is already fixed; ignoring the parameter.Nr[gMbP?r2)epsmomentum)rrr$ reduced_tail)@rcSsg|] }ddgqS)rr r+)rv_r+r+r,ry)rmz1ssdlite320_mobilenet_v3_large..g?gffffff?) min_ratio max_ratioz4The length of the output channels from the backbone z? do not match the length of the anchor generator aspect ratios g?i,)?rr) score_thresh nms_threshdetections_per_imgtopk_candidates image_mean image_stdhead)r)rverifyrwarningswarnrrrrrr BatchNorm2drr>rrrange det_utilsretrieve_out_channelsnum_anchors_per_location aspect_ratiosrrrr?load_state_dictget_state_dict)rrrBrrr$rZ reduce_tailreranchor_generatorr"rAdefaultsmodelr+r+r,r sp7       ) _ModelURLsZ"ssdlite320_mobilenet_v3_large_coco)@r collectionsr functoolsrtypingrrrrrr r7r r Zops.miscr Ztransforms._presetsrutilsrr_apirr_metarrrr mobilenetv3rrr anchor_utilsrbackbone_utilsrssdrr__all__rUrVr(r-r0r>r?rFrHrb MobileNetV2 MobileNetV3rrr IMAGENET1K_V1boolr rr model_urlsr+r+r+r,sv               1 y