o ={cV @sdZddlmmZddlmZddlmZddl m Z ddl m Z ddl mZddl mZdd lmZd Zd d d ddddddZda        dFddZdGddZdHd d!ZdId#d$ZdHd%d&Z   '  dJd(d)ZdKd*d+Zed,d-d.      dLd/d0Zed1d2      dLd3d4Zed5d6      dLd7d8Zed9d:dMd;d<Zed=d>dNd@dAZ ej!j"dBej#ej$dCe_ej je _dDZ%e&edEeje%e&edEeje%e&edEeje%dS)OzResNet models for Keras. Reference: - [Deep Residual Learning for Image Recognition]( https://arxiv.org/abs/1512.03385) (CVPR 2015) N)backend)imagenet_utils)training)VersionAwareLayers) data_utils) layer_utils) keras_exportzDhttps://storage.googleapis.com/tensorflow/keras-applications/resnet/)Z 2cb95161c43110f7111970584f804107Z 4d473c1dd8becc155b73f8504c6f6626)Z f1aeb4b969a6efcfb50fad2f0c20cfc5Z 88cf7a10940856eca736dc7b7e228a21)Z 100835be76be38e30d865e96f2aaae62Z ee4c566cf9a93f14d82f913c2dc6dd0c)Z 3ef43a0b657b3be2300d5770ece849e0Z fac2f116257151a9d068a22e544a4917)Z 6343647c601c52e1368623803854d971Z c0ed64b8031c3730f411d2eb4eea35b5)Z a49b44d1979771252814e80f8ec446f9Z ed17cf2e0169df9d443503ef94b23b33)Z 67a5b30d522ed92f75a1f16eef299d1aZ 62527c363bdd9ec598bed41947b379fc)Z 34fb605428fcc7aa4d62f44404c11509Z 0f678c91647380debd923963594981b3)resnet50 resnet101 resnet152Z resnet50v2Z resnet101v2Z resnet152v2Z resnext50Z resnext101resnetTimagenetsoftmaxc Ksd| vr | danta| rtd| |dvs%tjj|s%td|dkr3|r3| dkr3tdtj |dd t ||d }|d urKtj |d } nt |sXtj ||d } n|} t dkrbdnd} tjddd| }tjddd|dd|}|stj| ddd|}tjddd|}tjdd d|}tjddd!d"|}||}|rtj| dd#d|}tjdd$d|}|rtjd%d|}t| |tj| | d&d'|}n|d(krtjd%d|}n |d)krtjd*d|}|d urt|}n| }tj|||d}|dkr6|tvr6|r|d+}t|d,}n |d-}t|d}tj|t|d.|d/}|||S|d ur@|||S)0aInstantiates the ResNet, ResNetV2, and ResNeXt architecture. Args: stack_fn: a function that returns output tensor for the stacked residual blocks. preact: whether to use pre-activation or not (True for ResNetV2, False for ResNet and ResNeXt). use_bias: whether to use biases for convolutional layers or not (True for ResNet and ResNetV2, False for ResNeXt). model_name: string, model name. include_top: whether to include the fully-connected layer at the top of the network. weights: one of `None` (random initialization), 'imagenet' (pre-training on ImageNet), or the path to the weights file to be loaded. input_tensor: optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model. input_shape: optional shape tuple, only to be specified if `include_top` is False (otherwise the input shape has to be `(224, 224, 3)` (with `channels_last` data format) or `(3, 224, 224)` (with `channels_first` data format). It should have exactly 3 inputs channels. pooling: optional pooling mode for feature extraction when `include_top` is `False`. - `None` means that the output of the model will be the 4D tensor output of the last convolutional layer. - `avg` means that global average pooling will be applied to the output of the last convolutional layer, and thus the output of the model will be a 2D tensor. - `max` means that global max pooling will be applied. classes: optional number of classes to classify images into, only to be specified if `include_top` is True, and if no `weights` argument is specified. classifier_activation: A `str` or callable. The activation function to use on the "top" layer. Ignored unless `include_top=True`. Set `classifier_activation=None` to return the logits of the "top" layer. When loading pretrained weights, `classifier_activation` can only be `None` or `"softmax"`. **kwargs: For backwards compatibility only. Returns: A `keras.Model` instance. layerszUnknown argument(s): >r NzThe `weights` argument should be either `None` (random initialization), `imagenet` (pre-training on ImageNet), or the path to the weights file to be loaded.r rzWIf using `weights` as `"imagenet"` with `include_top` as true, `classes` should be 1000 )Z default_sizeZmin_size data_formatZrequire_flattenweightsN)shape)Ztensorr channels_last)rrrZ conv1_padZpaddingname@Z conv1_convstridesuse_biasr>Zconv1_bnZaxisepsilonrreluZ conv1_relurrrr(Z pool1_padZ pool1_poolr rZpost_bnZ post_reluZavg_poolZ predictions)Z activationrZavgmaxZmax_poolz&_weights_tf_dim_ordering_tf_kernels.h5rz,_weights_tf_dim_ordering_tf_kernels_notop.h5Zmodels)Z cache_subdir file_hash) poprr ValueErrortfioZgfileexistsrZobtain_input_shaperimage_data_formatZInputZis_keras_tensor ZeroPadding2DConv2DBatchNormalization Activation MaxPooling2DZGlobalAveragePooling2DZvalidate_activationZDenseZGlobalMaxPooling2DrZget_source_inputsrZModelWEIGHTS_HASHESrZget_fileBASE_WEIGHTS_PATHZ load_weights)stack_fnpreactr!Z model_name include_topr input_tensor input_shapepoolingclassesZclassifier_activationkwargsZ img_inputbn_axisxZinputsZmodel file_namer+Z weights_pathrD9lib/python3.10/site-packages/keras/applications/resnet.pyResNetLs>         rFrrcCsJtdkrdnd}|r(tjd|d||dd|}tj|d|dd |}n|}tj|d||d d|}tj|d|d d |}tjd |d d|}tj||d|dd|}tj|d|dd |}tjd |dd|}tjd|d|dd|}tj|d|dd |}tj|dd||g}tjd |dd|}|S)aA residual block. Args: x: input tensor. filters: integer, filters of the bottleneck layer. kernel_size: default 3, kernel size of the bottleneck layer. stride: default 1, stride of the first layer. conv_shortcut: default True, use convolution shortcut if True, otherwise identity shortcut. name: string, block label. Returns: Output tensor for the residual block. rrr_0_convr)r"_0_bnr#_1_conv_1_bnr%_1_relur&ZSAME_2_convr_2_bn_2_relu_3_conv_3_bn_add_out)rr1rr3r4r5Add)rBfilters kernel_sizestride conv_shortcutrrAshortcutrDrDrEblock1sP     rZrcCsHt||||dd}td|dD]}t||d|dt|d}q|S)adA set of stacked residual blocks. Args: x: input tensor. filters: integer, filters of the bottleneck layer in a block. blocks: integer, blocks in the stacked blocks. stride1: default 2, stride of the first layer in the first block. name: string, stack label. Returns: Output tensor for the stacked blocks. _block1rWrrrF_blockrXr)rZrangestrrBrUblocksstride1rirDrDrEstack1(s  reFc Csftdkrdnd}tj|d|dd|}tjd|dd |}|r3tjd |d||d d |}n|dkr@tjd|d |n|}tj|ddd|dd|}tj|d|dd|}tjd|dd |}tjd|dd|}tj|||d|dd|}tj|d|dd|}tjd|dd |}tjd |d|dd |}tj|dd ||g}|S)aA residual block. Args: x: input tensor. filters: integer, filters of the bottleneck layer. kernel_size: default 3, kernel size of the bottleneck layer. stride: default 1, stride of the first layer. conv_shortcut: default False, use convolution shortcut if True, otherwise identity shortcut. name: string, block label. Returns: Output tensor for the residual block. rrrr"Z _preact_bnr#r%Z _preact_relur&rGrHr))r FrJrrKrLr'_2_padrrMrNrOrPrS) rr1rr4r5r3r6r2rT) rBrUrVrWrXrrAr:rYrDrDrEblock2=sZ   rgcCs^t||d|dd}td|D]}t|||dt|d}qt||||dt|d}|S)apA set of stacked residual blocks. Args: x: input tensor. filters: integer, filters of the bottleneck layer in a block. blocks: integer, blocks in the stacked blocks. stride1: default 2, stride of the first layer in the first block. name: string, stack label. Returns: Output tensor for the stacked blocks. Tr[r^rr]r&r\)rgr_r`rarDrDrEstack2vs  rhrc stdkrdnd}|r+tjd||d|d|dd|}tj|d|d d |}n|}tj|dd|d d |}tj|d|d d |}tjd|dd|}||tjd|dd|}tj||d|dd|}t|dd} t |t | |fg}tj fdd|dd|}t |t | |fg}tj|d|dd |}tjd|dd|}tjd||dd|dd |}tj|d|dd |}tj |dd||g}tjd|d d|}|S)!aA residual block. Args: x: input tensor. filters: integer, filters of the bottleneck layer. kernel_size: default 3, kernel size of the bottleneck layer. stride: default 1, stride of the first layer. groups: default 32, group size for grouped convolution. conv_shortcut: default True, use convolution shortcut if True, otherwise identity shortcut. name: string, block label. Returns: Output tensor for the residual block. rrrrFrHrr"rIr#rJ)r!rrKr%rLr&r'rfrrM)r Zdepth_multiplierr!rNcstfddtDS)Nc3s2|]}dddddddd|fVqdSNrD).0rdrBrDrE s0z+block3....)sumr_rlcrlrEszblock3..Z _2_reducerNrOrPrQrRrS) rr1rr3r4r5r2ZDepthwiseConv2DrZreshapeZ concatenateZLambdarT) rBrUrVrWgroupsrXrrArYZx_shaperDrorEblock3s~      rsc CsLt|||||dd}td|dD]}t|||d|dt|d}q|S)aA set of stacked residual blocks. Args: x: input tensor. filters: integer, filters of the bottleneck layer in a block. blocks: integer, blocks in the stacked blocks. stride1: default 2, stride of the first layer in the first block. groups: default 32, group size for grouped convolution. name: string, stack label. Returns: Output tensor for the stacked blocks. r[)rWrrrrrFr])rrrXr)rsr_r`)rBrUrbrcrrrrdrDrDrEstack3srtz$keras.applications.resnet50.ResNet50z"keras.applications.resnet.ResNet50zkeras.applications.ResNet50c K*dd}t|ddd||||||f i|S)z'Instantiates the ResNet50 architecture.cSBt|ddddd}t|dddd }t|d d d d }t|d ddd S)Nrrrconv2rcrrGconv3r&conv4conv5rerlrDrDrEr9zResNet50..stack_fnFTr rFr;rr<r=r>r?r@r9rDrDrEResNet50s rz#keras.applications.resnet.ResNet101zkeras.applications.ResNet101c Kru)z(Instantiates the ResNet101 architecture.cSrv)NrrrrwrxryrGrzr&r{r}r~rrrlrDrDrEr9&rzResNet101..stack_fnFTr rrrDrDrE ResNet101 rz#keras.applications.resnet.ResNet152zkeras.applications.ResNet152c Kru)z(Instantiates the ResNet152 architecture.cSrv)Nrrrrwrxryrzr&r{$r}r~rrrlrDrDrEr9IrzResNet152..stack_fnFTr rrrDrDrE ResNet152;rrz,keras.applications.resnet50.preprocess_inputz*keras.applications.resnet.preprocess_inputcCstj||ddS)NZcaffe)rmode)rpreprocess_input)rBrrDrDrEr^srz.keras.applications.resnet50.decode_predictionsz,keras.applications.resnet.decode_predictionscCstj||dS)N)top)rdecode_predictions)ZpredsrrDrDrErhsr)rreterrora9 Reference: - [Deep Residual Learning for Image Recognition]( https://arxiv.org/abs/1512.03385) (CVPR 2015) For image classification use cases, see [this page for detailed examples]( https://keras.io/api/applications/#usage-examples-for-image-classification-models). For transfer learning use cases, make sure to read the [guide to transfer learning & fine-tuning]( https://keras.io/guides/transfer_learning/). Note: each Keras Application expects a specific kind of input preprocessing. For ResNet, call `tf.keras.applications.resnet.preprocess_input` on your inputs before passing them to the model. `resnet.preprocess_input` will convert the input images from RGB to BGR, then will zero-center each color channel with respect to the ImageNet dataset, without scaling. Args: include_top: whether to include the fully-connected layer at the top of the network. weights: one of `None` (random initialization), 'imagenet' (pre-training on ImageNet), or the path to the weights file to be loaded. input_tensor: optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model. input_shape: optional shape tuple, only to be specified if `include_top` is False (otherwise the input shape has to be `(224, 224, 3)` (with `'channels_last'` data format) or `(3, 224, 224)` (with `'channels_first'` data format). It should have exactly 3 inputs channels, and width and height should be no smaller than 32. E.g. `(200, 200, 3)` would be one valid value. pooling: Optional pooling mode for feature extraction when `include_top` is `False`. - `None` means that the output of the model will be the 4D tensor output of the last convolutional block. - `avg` means that global average pooling will be applied to the output of the last convolutional block, and thus the output of the model will be a 2D tensor. - `max` means that global max pooling will be applied. classes: optional number of classes to classify images into, only to be specified if `include_top` is True, and if no `weights` argument is specified. classifier_activation: A `str` or callable. The activation function to use on the "top" layer. Ignored unless `include_top=True`. Set `classifier_activation=None` to return the logits of the "top" layer. When loading pretrained weights, `classifier_activation` can only be `None` or `"softmax"`. Returns: A Keras model instance. __doc__)r Tr NNNrr)rrTN)rN)rrFN)rrrTN)rrN)Tr NNNrrj)r)'rZtensorflow.compat.v2compatZv2r.ZkerasrZkeras.applicationsrZ keras.enginerZ keras.layersrZ keras.utilsrrZ tensorflow.python.util.tf_exportrr8r7rrFrZrergrhrsrtrrrrrZPREPROCESS_INPUT_DOCformatZPREPROCESS_INPUT_RET_DOC_CAFFEZPREPROCESS_INPUT_ERROR_DOCZDOCsetattrrDrDrDrEs       #  * 3  9 O         <