\;j(ddlZddlmZddlmZgZdddZGddej Zdd Zgd gd gd gd dZ dZ ddZ ddZ ddZ ddZy)N)nn)get_weights_path_from_url)z7https://paddle-hapi.bj.bcebos.com/models/vgg16.pdparams 89bbffc0f87d260be9b8cdc169c991c4)z7https://paddle-hapi.bj.bcebos.com/models/vgg19.pdparams 23b18bb13d8894f60f54e642be79a0dd)vgg16vgg19c*eZdZdZdfd ZdZxZS)VGGa4VGG model from `"Very Deep Convolutional Networks For Large-Scale Image Recognition" `_. Args: features (nn.Layer): Vgg features create by function make_layers. 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 three fc layer or not. Default: True. Returns: :ref:`api_paddle_nn_Layer`. An instance of VGG model. Examples: .. code-block:: python >>> import paddle >>> from paddle.vision.models import VGG >>> from paddle.vision.models.vgg import make_layers >>> vgg11_cfg = [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M'] >>> features = make_layers(vgg11_cfg) >>> vgg11 = VGG(features) >>> x = paddle.rand([1, 3, 224, 224]) >>> out = vgg11(x) >>> print(out.shape) [1, 1000] c t|||_||_||_|rt j d|_|dkDrt jt jddt jt jt jddt jt jt jd||_ yy)N)r ribi) super__init__features num_classes with_poolrAdaptiveAvgPool2Davgpool SequentialLinearReLUDropout classifier)selfrrr __class__s aG:\00. PROJECTS\API\Inventory\templateJSON\kerjaOCR\Lib\site-packages\paddle/vision/models/vgg.pyrz VGG.__init__Bs   &" //7DL ? mm +t,   $%   $ ,DO c|j|}|jr|j|}|jdkDr't j |d}|j |}|S)Nr)rrrrpaddleflattenr)rxs rforwardz VGG.forwardVsU MM!  >> QA   a q!$A"Ar)iT)__name__ __module__ __qualname____doc__rr" __classcell__)rs@rr r !s@( rr cFg}d}|D]}|dk(r|tjddgz }$tj||dd}|r.||tj|tjgz }n||tjgz }|}tj |S)NM) kernel_sizestrider)r,padding)r MaxPool2DConv2D BatchNorm2Drr)cfg batch_normlayers in_channelsvconv2ds r make_layersr8cs FK  8 r||!<= =FYY{A1aHF62>>!#4bggi@@62779--K ==& !!r) @r*r*r;r*r<r*r<r<r*)r9r9r*r:r:r*r;r;r*r<r<r*r<r<r*)r9r9r*r:r:r*r;r;r;r*r<r<r<r*r<r<r<r*)r9r9r*r:r:r*r;r;r;r;r*r<r<r<r<r*r<r<r<r<r*)ABDEc ttt||fi|}|rX|tvs J|dt t|dt|d}t j |}|j||S)N)r3zJ model do not have a pretrained model now, you should set pretrained=Falserr)r r8cfgs model_urlsrrload load_dict)archr2r3 pretrainedkwargsmodel weight_pathparams r_vggrLs  DI*= H HE J  _V] ^ _ / t Q D!1!!4  K(  Lrc 2d}|r|dz }t|d||fi|S)a0VGG 11-layer model from `"Very Deep Convolutional Networks For Large-Scale Image Recognition" `_. Args: pretrained (bool, optional): Whether to load pre-trained weights. If True, returns a model pre-trained on ImageNet. Default: False. batch_norm (bool, optional): If True, returns a model with batch_norm layer. Default: False. **kwargs (optional): Additional keyword arguments. For details, please refer to :ref:`VGG `. Returns: :ref:`api_paddle_nn_Layer`. An instance of VGG 11-layer model. Examples: .. code-block:: python >>> import paddle >>> from paddle.vision.models import vgg11 >>> # build model >>> model = vgg11() >>> # build vgg11 model with batch_norm >>> model = vgg11(batch_norm=True) >>> x = paddle.rand([1, 3, 224, 224]) >>> out = model(x) >>> print(out.shape) [1, 1000] vgg11_bnr=rLrGr3rH model_names rrNrN,>Je  CZ B6 BBrc 2d}|r|dz }t|d||fi|S)a&VGG 13-layer model from `"Very Deep Convolutional Networks For Large-Scale Image Recognition" `_. Args: pretrained (bool, optional): Whether to load pre-trained weights. If True, returns a model pre-trained on ImageNet. Default: False. batch_norm (bool): If True, returns a model with batch_norm layer. Default: False. **kwargs (optional): Additional keyword arguments. For details, please refer to :ref:`VGG `. Returns: :ref:`api_paddle_nn_Layer`. An instance of VGG 13-layer model. Examples: .. code-block:: python >>> import paddle >>> from paddle.vision.models import vgg13 >>> # build model >>> model = vgg13() >>> # build vgg13 model with batch_norm >>> model = vgg13(batch_norm=True) >>> x = paddle.rand([1, 3, 224, 224]) >>> out = model(x) >>> print(out.shape) [1, 1000] vgg13rOr>rPrQs rrUrUrSrc 2d}|r|dz }t|d||fi|S)a0VGG 16-layer model from `"Very Deep Convolutional Networks For Large-Scale Image Recognition" `_. Args: pretrained (bool, optional): Whether to load pre-trained weights. If True, returns a model pre-trained on ImageNet. Default: False. batch_norm (bool, optional): If True, returns a model with batch_norm layer. Default: False. **kwargs (optional): Additional keyword arguments. For details, please refer to :ref:`VGG `. Returns: :ref:`api_paddle_nn_Layer`. An instance of VGG 16-layer model. Examples: .. code-block:: python >>> import paddle >>> from paddle.vision.models import vgg16 >>> # build model >>> model = vgg16() >>> # build vgg16 model with batch_norm >>> model = vgg16(batch_norm=True) >>> x = paddle.rand([1, 3, 224, 224]) >>> out = model(x) >>> print(out.shape) [1, 1000] rrOr?rPrQs rrrrSrc 2d}|r|dz }t|d||fi|S)a0VGG 19-layer model from `"Very Deep Convolutional Networks For Large-Scale Image Recognition" `_. Args: pretrained (bool, optional): Whether to load pre-trained weights. If True, returns a model pre-trained on ImageNet. Default: False. batch_norm (bool, optional): If True, returns a model with batch_norm layer. Default: False. **kwargs (optional): Additional keyword arguments. For details, please refer to :ref:`VGG `. Returns: :ref:`api_paddle_nn_Layer`. An instance of VGG 19-layer model. Examples: .. code-block:: python >>> import paddle >>> from paddle.vision.models import vgg19 >>> # build model >>> model = vgg19() >>> # build vgg19 model with batch_norm >>> model = vgg19(batch_norm=True) >>> x = paddle.rand([1, 3, 224, 224]) >>> out = model(x) >>> print(out.shape) [1, 1000] rrOr@rPrQs rrr4rSr)F)FF)rrpaddle.utils.downloadr__all__rCLayerr r8rBrLrNrUrrrrr\sz;   ?"((?D "" J " ( O>B""CJ"CJ"CJ"Cr