\;j%ddlZddlmcmZddlmZddlmZddlmZm Z m Z m Z ddl m Z gZdddZGd d ej ZGd d ej ZGd dej ZdZddZddZy)N)nn) ParamAttr)AdaptiveAvgPool2DConv2DDropout MaxPool2D)get_weights_path_from_url)z[https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SqueezeNet1_0_pretrained.pdparams 30b95af60a2178f03cf9b66cd77e1db1)z[https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SqueezeNet1_1_pretrained.pdparams a11250d3a1f91d7131fd095ebbf09eee) squeezenet1_0 squeezenet1_1c&eZdZdfd ZdZxZS) MakeFireConvc nt|t||||tt|_y)N)padding weight_attr bias_attr)super__init__rr_conv)selfinput_channelsoutput_channels filter_sizer __class__s hG:\00. PROJECTS\API\Inventory\templateJSON\kerjaOCR\Lib\site-packages\paddle/vision/models/squeezenet.pyrzMakeFireConv.__init__%s0    ! k   cR|j|}tj|}|S)N)rFrelu)rxs rforwardzMakeFireConv.forward0s! JJqM FF1Ir)r__name__ __module__ __qualname__rr" __classcell__rs@rrr$s   rrc$eZdZfdZdZxZS)MakeFirect|t||d|_t||d|_t||dd|_y)N)r)rrrr _conv_path1 _conv_path2)rrsqueeze_channelsexpand1x1_channelsexpand3x3_channelsrs rrzMakeFire.__init__7sK !.2BAF '(8:LaP' 0!Q rc|j|}|j|}|j|}tj||gdS)Nr,axis)rr.r/paddleconcat)rinputsr!x1x2s rr"zMakeFire.forwardEsE JJv    a    a }}b"XA..rr#r(s@rr*r*6s   /rr*c*eZdZdZdfd ZdZxZS) SqueezeNetaSqueezeNet model from `"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size" `_. Args: version (str): Version of SqueezeNet, which can be "1.0" or "1.1". num_classes (int, optional): Output dim of last fc layer. If num_classes <= 0, last fc layer will not be defined. Default: 1000. with_pool (bool, optional): Use pool before the last fc layer or not. Default: True. Returns: :ref:`api_paddle_nn_Layer`. An instance of SqueezeNet model. Examples: .. code-block:: python >>> import paddle >>> from paddle.vision.models import SqueezeNet >>> # build v1.0 model >>> model = SqueezeNet(version='1.0') >>> # build v1.1 model >>> # model = SqueezeNet(version='1.1') >>> x = paddle.rand([1, 3, 224, 224]) >>> out = model(x) >>> print(out.shape) [1, 1000] c jt|||_||_||_ddg}||vs Jd|d||jdk(rt ddddt t |_tddd |_ tdd d d |_ tdd d d |_ tdddd|_ tdddd|_tdddd|_tdddd|_tdd dd|_tdd dd|_nt dd dddt t |_tddd |_ td d d d |_ tdd d d |_ tdddd|_ tdddd|_tdddd|_tdddd|_tdd dd|_tdd dd|_t'dd|_t d|dt t |_t-d|_y)N1.01.1zsupported versions are z but input version is r-`)striderrr) kernel_sizerCr@ 0iir,)rCrrrg?downscale_in_infer)pmode)rr)rrversion num_classes with_poolrrrr_poolr*_conv1_conv2_conv3_conv4_conv5_conv6_conv7_conv8r_drop_conv9r _avg_pool)rrOrPrQsupported_versionsrs rrzSqueezeNet.__init__ms*  &"#U^ ) ) Y $%7$88Nwi X Y ) <<5 %K#+ DJ#qAFDJ"2r2r2DK"3B3DK"3C5DK"3C5DK"3C5DK"3C5DK"3C5DK"3C5DK%K#+DJ#qAFDJ"2r2r2DK"3B3DK"3C5DK"3C5DK"3C5DK"3C5DK"3C5DK"3C5DKs)=>  aY[IK  +1-rc4|j|}tj|}|j|}|jdk(r|j |}|j |}|j|}|j|}|j|}|j|}|j|}|j|}|j|}|j|}n|j |}|j |}|j|}|j|}|j|}|j|}|j|}|j|}|j|}|j|}|jdkDr"|j|}|j|}|j r?tj|}|j#|}t%j&|ddg}|S)Nr>rrBr-r4)rrr rRrOrSrTrUrVrWrXrYrZrPr[r\rQr]r6squeeze)rr8r!s rr"zSqueezeNet.forwards JJv  FF1I JJqM <<5  AA AA AA 1 A AA AA AA AA 1 A AA AA AA 1 A AA AA 1 A AA AA AA AA   a  1 A AA >>q Aq!Aq1v.Ar)iT)r$r%r&__doc__rr"r'r(s@rr<r<Ls@5.n"rr<c t|fi|}|rX|tvs J|dtt|dt|d}tj|}|j ||S)NzJ model do not have a pretrained model now, you should set pretrained=Falserr,)r< model_urlsr r6loadset_dict)archrO pretrainedkwargsmodel weight_pathparams r _squeezenetrlsz w )& )E J  _V] ^ _ / t Q D!1!!4  K( u Lrc tdd|fi|S)aSqueezeNet v1.0 model from `"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size" `_. Args: pretrained (bool, optional): Whether to load pre-trained weights. If True, returns a model pre-trained on ImageNet. Default: False. **kwargs (optional): Additional keyword arguments. For details, please refer to :ref:`SqueezeNet `. Returns: :ref:`api_paddle_nn_Layer`. An instance of SqueezeNet v1.0 model. Examples: .. code-block:: python >>> import paddle >>> from paddle.vision.models import squeezenet1_0 >>> # build model >>> model = squeezenet1_0() >>> # build model and load imagenet pretrained weight >>> # model = squeezenet1_0(pretrained=True) >>> x = paddle.rand([1, 3, 224, 224]) >>> out = model(x) >>> print(out.shape) [1, 1000] r r>rlrgrhs rr r > z DV DDrc tdd|fi|S)aSqueezeNet v1.1 model from `"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size" `_. Args: pretrained (bool, optional): Whether to load pre-trained weights. If True, returns a model pre-trained on ImageNet. Default: False. **kwargs (optional): Additional keyword arguments. For details, please refer to :ref:`SqueezeNet `. Returns: :ref:`api_paddle_nn_Layer`. An instance of SqueezeNet v1.1 model. Examples: .. code-block:: python >>> import paddle >>> from paddle.vision.models import squeezenet1_1 >>> # build model >>> model = squeezenet1_1() >>> # build model and load imagenet pretrained weight >>> # model = squeezenet1_1(pretrained=True) >>> x = paddle.rand([1, 3, 224, 224]) >>> out = model(x) >>> print(out.shape) [1, 1000] r r?rnros rr r rpr)F)r6paddle.nn.functionalr functionalrpaddle.base.param_attrr paddle.nnrrrrpaddle.utils.downloadr __all__rcLayerrr*r<rlr r rrrzsz ,CC;   288$/rxx/,zzz EDEr