hnBbddlmZddlmZmZmZddlZddlmZddlm Z ddl m Z ddl m Z dd lmZmZmZdd lmZdd lmZmZdd l mZmZmZmZd dlmZmZmZgdZGdde j@Z!Gdde jDZ#de$e%de$e%deede&de&dede#fdZ'dedddd Z(Gd!d"eZ)Gd#d$eZ*Gd%d&eZ+Gd'd(eZ,ed)*ed+d,f-dd.d/d0deee)efde&de&dede#f d1Z-ed2*ed+d3f-dd.d/d0deee*efde&de&dede#f d4Z.ed5*ed+d6f-dd.d/d0deee+efde&de&dede#f d7Z/ed8*ed+d9f-dd.d/d0deee,efde&de&dede#f d:Z0y);)partial)AnyOptionalUnionN)Tensor) shufflenetv2)ImageClassification)register_modelWeights WeightsEnum)_IMAGENET_CATEGORIES)_ovewrite_named_paramhandle_legacy_interface)ShuffleNet_V2_X0_5_WeightsShuffleNet_V2_X1_0_WeightsShuffleNet_V2_X1_5_WeightsShuffleNet_V2_X2_0_Weights) _fuse_modules _replace_reluquantize_model) QuantizableShuffleNetV2#ShuffleNet_V2_X0_5_QuantizedWeights#ShuffleNet_V2_X1_0_QuantizedWeights#ShuffleNet_V2_X1_5_QuantizedWeights#ShuffleNet_V2_X2_0_QuantizedWeightsshufflenet_v2_x0_5shufflenet_v2_x1_0shufflenet_v2_x1_5shufflenet_v2_x2_0c<eZdZdededdffd ZdedefdZxZS)QuantizableInvertedResidualargskwargsreturnNcjt||i|tjj |_yN)super__init__nn quantizedFloatFunctionalcatselfr%r& __class__s j/var/www/html/eduruby.in/venv/lib/python3.12/site-packages/torchvision/models/quantization/shufflenetv2.pyr+z$QuantizableInvertedResidual.__init__$s) $)&)<<//1xcT|jdk(rE|jdd\}}|jj||j|gd}n=|jj|j ||j|gd}t j |d}|S)Nrr )dim)stridechunkr/branch2branch1rchannel_shuffle)r1r5x1x2outs r3forwardz#QuantizableInvertedResidual.forward(s ;;! WWQAW&FB((,,DLL$451,=C((,, QaAq,IC**32 r4)__name__ __module__ __qualname__rr+rr@ __classcell__r2s@r3r$r$#s02c2S2T2  F r4r$cTeZdZdededdffd ZdedefdZd deeddfd Z xZ S) rr%r&r'Nct||dti|tjj j |_tjj j|_ y)Ninverted_residual) r*r+r$torchao quantization QuantStubquant DeQuantStubdequantr0s r3r+z QuantizableShuffleNetV2.__init__6sP $X2MXQWXXX**446 xx,,88: r4r5cl|j|}|j|}|j|}|Sr))rM _forward_implrO)r1r5s r3r@zQuantizableShuffleNetV2.forward;s1 JJqM   q ! LLOr4is_qatc|jjD]\}}|dvs |t|gdg|d!|jD]}t |t ust |jjjdkDrt|jddggd g|dt|jgdd d ggd g|dy) aBFuse conv/bn/relu modules in shufflenetv2 model Fuse conv+bn+relu/ conv+relu/conv+bn modules to prepare for quantization. Model is modified in place. .. note:: Note that this operation does not change numerics and the model after modification is in floating point )conv1conv5N)012T)inplacerrVrW)rX34rZr[)567) _modulesitemsrmodulestyper$lenr;r:)r1rRnamems r3 fuse_modelz"QuantizableShuffleNetV2.fuse_modelAs}}**, JGD!))ama/!2FDI J AAw55qyy))//12Q6!!))sCj/-JF\`aII$sCj/B  r4r)) rArBrCrr+rr@rboolrfrDrEs@r3rr4sG;c;S;T; F $4r4rstages_repeatsstages_out_channelsweightsprogressquantizer&r'c X|Kt|dt|jdd|jvrt|d|jd|jdd}t ||fi|}t ||r t |||"|j|j|d|S)N num_classes categoriesbackendfbgemmT)rk check_hash) rrcmetapoprrrload_state_dictget_state_dict)rhrirjrkrlr&rpmodels r3 _shufflenetv2rxZsfmSl9S5TU  $ !&)W\\)5L MjjH-G #N4G R6 RE%ug& g44hSW4XY Lr4)rrrqzdhttps://github.com/pytorch/vision/tree/main/references/classification#post-training-quantized-modelsz These weights were produced by doing Post Training Quantization (eager mode) on top of the unquantized weights listed below. )min_sizerorprecipe_docsc heZdZedeediedejddddid d d  Z e Z y )rzShttps://download.pytorch.org/models/quantized/shufflenetv2_x0.5_fbgemm-00845098.pth crop_sizei ImageNet-1Kg#~jL@gRS@zacc@1zacc@5g{Gz?gjt? num_params unquantized_metrics_ops _file_sizeurl transformsrsN) rArBrCr rr _COMMON_METAr IMAGENET1K_V1IMAGENET1K_FBGEMM_V1DEFAULTr4r3rrs[" a.#>    !5CC##    "#Gr4rc heZdZedeediedejddddid d d  Z e Z y )rzQhttps://download.pytorch.org/models/quantized/shufflenetv2_x1_fbgemm-1e62bb32.pthr}r~i"rgףp= Q@gh|?U@rg(\?gy&1@rrN) rArBrCr rr rrrrrrr4r3rrs[" _.#>    !5CC##    "#Gr4rc leZdZedeeddieddejddd d id d d Z e Z y)rzShttps://download.pytorch.org/models/quantized/shufflenetv2_x1_5_fbgemm-d7401f05.pthr}r resize_size+https://github.com/pytorch/vision/pull/5906iv5rgSR@g̬V@rgl?gK7A` @rzrrrrrrN) rArBrCr rr rrrrrrr4r3rr`" a.#3O    C!5CC##    $#Gr4rc leZdZedeeddieddejddd d id d d Z e Z y)rzShttps://download.pytorch.org/models/quantized/shufflenetv2_x2_0_fbgemm-5cac526c.pthr}rrriprg-R@gZd;W@rg-?g|?5@rrN) rArBrCr rr rrrrrrr4r3rrrr4rquantized_shufflenet_v2_x0_5)rd pretrainedcf|jddrtjStjSNrlF)getrrrrr&s r3r0zz*e, 0 D D,99r4)rjTFrjrkrlc f|rtntj|}tgdgdf|||d|S)aQ Constructs a ShuffleNetV2 with 0.5x output channels, as described in `ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design `__. .. 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.ShuffleNet_V2_X0_5_QuantizedWeights` or :class:`~torchvision.models.ShuffleNet_V2_X0_5_Weights`, optional): The pretrained weights for the model. See :class:`~torchvision.models.quantization.ShuffleNet_V2_X0_5_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.ShuffleNet_V2_X0_5_QuantizedWeights`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.quantization.ShuffleNet_V2_X0_5_QuantizedWeights :members: .. autoclass:: torchvision.models.ShuffleNet_V2_X0_5_Weights :members: :noindex: r)0`r)rrverifyrxrjrkrlr&s r3rrsDd7?2D^ffgnoG * 4;hYa ek r4quantized_shufflenet_v2_x1_0cf|jddrtjStjSr)rrrrrrs r3rrrr4c f|rtntj|}tgdgdf|||d|S)aQ Constructs a ShuffleNetV2 with 1.0x output channels, as described in `ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design `__. .. 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.ShuffleNet_V2_X1_0_QuantizedWeights` or :class:`~torchvision.models.ShuffleNet_V2_X1_0_Weights`, optional): The pretrained weights for the model. See :class:`~torchvision.models.quantization.ShuffleNet_V2_X1_0_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.ShuffleNet_V2_X1_0_QuantizedWeights`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.quantization.ShuffleNet_V2_X1_0_QuantizedWeights :members: .. autoclass:: torchvision.models.ShuffleNet_V2_X1_0_Weights :members: :noindex: r)rtrirr)rrrrxrs r3r r Dd7?2D^ffgnoG , 6=[c gm r4quantized_shufflenet_v2_x1_5cf|jddrtjStjSr)rrrrrrs r3rrJrr4c f|rtntj|}tgdgdf|||d|S)aQ Constructs a ShuffleNetV2 with 1.5x output channels, as described in `ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design `__. .. 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.ShuffleNet_V2_X1_5_QuantizedWeights` or :class:`~torchvision.models.ShuffleNet_V2_X1_5_Weights`, optional): The pretrained weights for the model. See :class:`~torchvision.models.quantization.ShuffleNet_V2_X1_5_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.ShuffleNet_V2_X1_5_QuantizedWeights`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.quantization.ShuffleNet_V2_X1_5_QuantizedWeights :members: .. autoclass:: torchvision.models.ShuffleNet_V2_X1_5_Weights :members: :noindex: r)ri`irr)rrrrxrs r3r!r!Frr4quantized_shufflenet_v2_x2_0cf|jddrtjStjSr)rrrrrrs r3rrrr4c f|rtntj|}tgdgdf|||d|S)aQ Constructs a ShuffleNetV2 with 2.0x output channels, as described in `ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design `__. .. 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.ShuffleNet_V2_X2_0_QuantizedWeights` or :class:`~torchvision.models.ShuffleNet_V2_X2_0_Weights`, optional): The pretrained weights for the model. See :class:`~torchvision.models.quantization.ShuffleNet_V2_X2_0_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.ShuffleNet_V2_X2_0_QuantizedWeights`` base class. Please refer to the `source code `_ for more details about this class. .. autoclass:: torchvision.models.quantization.ShuffleNet_V2_X2_0_QuantizedWeights :members: .. autoclass:: torchvision.models.ShuffleNet_V2_X2_0_Weights :members: :noindex: r)riiir)rrrrxrs r3r"r"~rr4)1 functoolsrtypingrrrrItorch.nnr,rtorchvision.modelsrtransforms._presetsr _apir r r_metar_utilsrrrrrrutilsrrr__all__InvertedResidualr$ ShuffleNetV2rlistintrgrxrrrrrrr r!r"rr4r3rs'' +677(C @? ,"?"?"#l77#LIck "     6&t  #+#*#+#*#+#,#+#,34   ae * e?A[[\ ]** *  *  * 5*Z34   ae * e?A[[\ ]** *  *  * 5*Z34   ae * e?A[[\ ]** *  *  * 5*Z34   ae * e?A[[\ ]** *  *  * 5*r4