2VhBddlmZddlmZddlmZddlmZddlmZddl m Z ddl m Z dZ e d zZe d zZe d zZe d zZe d zZe dzZdZdZdZ d%dZeddg d&dZeddg d'dZeddg d(dZedd)dZedd*d Zej:j=d!ej>ej@"e_!ej8jBe_!d#Z"e#ed$ejBe"ze#ed$ejBe"ze#ed$ejBe"zy)+)backend)layers) keras_export)imagenet_utils) Functional)operation_utils) file_utilszFhttps://storage.googleapis.com/tensorflow/keras-applications/densenet/1densenet121_weights_tf_dim_ordering_tf_kernels.h57densenet121_weights_tf_dim_ordering_tf_kernels_notop.h51densenet169_weights_tf_dim_ordering_tf_kernels.h57densenet169_weights_tf_dim_ordering_tf_kernels_notop.h51densenet201_weights_tf_dim_ordering_tf_kernels.h57densenet201_weights_tf_dim_ordering_tf_kernels_notop.h5c ft|D]"}t|d|dzt|dzz}$|S)zA dense block. Args: x: input tensor. blocks: integer, the number of building blocks. name: string, block label. Returns: Output tensor for the block. _blockname)range conv_blockstr)xblocksris O/home/dcms/DCMS/lib/python3.12/site-packages/keras/src/applications/densenet.py dense_blockr#s=6]A q"4(?SQZ#? @A Hcvtjdk(rdnd}tj|d|dz|}tjd|dz |}tj t |j||zdd |d z |}tjd d |dz|}|S)zA transition block. Args: x: input tensor. reduction: float, compression rate at transition layers. name: string, block label. Returns: Output tensor for the block. channels_lastr>_bnaxisepsilonrrelu_relurF_convuse_biasr_poolstridesr) rimage_data_formatrBatchNormalization ActivationConv2DintshapeAveragePooling2D)r reductionrbn_axiss rtransition_blockr93s,,./AaqG !! hTE\  A 7&tg~6q9A   AGGG y ()  G^     A C14'>B1EA Hrc tjdk(rdnd}tj|d|dz|}tjd|dz |}tj d |zdd |d z |}tj|d|dz|}tjd|dz |}tj |ddd |dz|}tj ||dz||g}|S)zA building block for a dense block. Args: x: input tensor. growth_rate: float, growth rate at dense layers. name: string, block label. Returns: Output tensor for the block. r r!rr"_0_bnr$r'_0_relurF_1_convr*_1_bn_1_relusame_2_conv)paddingr+r_concat)r%r)rr0rr1r2r3 Concatenate)r growth_raterr8x1s rrrMs),,./AaqG  " " hTG^  B :  6y(8 9" =B  KU 1A  B  " " hTG^  B :  6y(8 9" =B QTI=M  B @dY.>?BHA HrNc Ttjdk(r td|dvs tj|s td|dk(r|r|dk7r tdt j |dd tj|| }|tj| } n/tj|stj|| } n|} tjd k(rdnd} tjd| } tjddddd| } tj| dd| } tjdd| } tjd| } tjddd | } t| |d!d"} t!| d#d$} t| |dd%} t!| d#d&} t| |dd'} t!| d#d(} t| |dd)} tj| dd*| } tjdd| } |rQtj"d+| } t j$||tj&||d,-| } nC|d.k(rtj"d+| } n!|d/k(rtj(d0| } |t+j,|} n| } t/| | |} |dk(r |rz|gd1k(rtj0d2t2d3d45}n|gd6k(rtj0d7t4d3d85}n|gd9k(rtj0d:t6d3d;5}ntd<|gd1k(rtj0d=t8d3d>5}nU|gd6k(rtj0d?t:d3d@5}n0|gd9k(rtj0dAt<d3dB5}n td<| j?|| S|| j?|| S)CaO Instantiates the DenseNet architecture. Reference: - [Densely Connected Convolutional Networks]( https://arxiv.org/abs/1608.06993) (CVPR 2017) This function returns a Keras image classification model, optionally loaded with weights pre-trained on ImageNet. 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 DenseNet, call `keras.applications.densenet.preprocess_input` on your inputs before passing them to the model. `densenet.preprocess_input` will scale pixels between 0 and 1 and then will normalize each channel with respect to the ImageNet dataset statistics. Args: blocks: numbers of building blocks for the four dense layers. 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. Defaults to `1000`. 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"`. name: The name of the model (string). Returns: A model instance. channels_firstzDenseNet does not support the `channels_first` image data format. Switch to `channels_last` by editing your local config file at ~/.keras/keras.json>NimagenetzThe `weights` argument should be either `None` (random initialization), `imagenet` (pre-training on ImageNet), or the path to the weights file to be loaded.rJzWIf using `weights` as `"imagenet"` with `include_top` as true, `classes` should be 1000r) default_sizemin_size data_formatrequire_flattenweights)r5)tensorr5r r!r)r!r!rS)rC@r,F conv1_conv)r/r+rr"conv1_bnr$r' conv1_relur)rrrYpool1r.rconv2g?pool2conv3pool3conv4pool4conv5bnavg_pool predictions) activationravgmaxmax_pool r models 9d60b8095a5708f2dcce2bca79d332c7) cache_subdir file_hashrjrkrrr d699b8f76981ab1b30698df4c175e90brjrk0rr 1ceb130c1ea1b78c3bf6114dbdfd8807zweights_path undefinedr 30ee3e1110167f948a6b9946edeeb738r b8c4d4c20dd625c148057b9ff1c1176br c13680b51ded0fb44dff2d8f86ac8bb1) rr0 ValueErrorr existsrobtain_input_shaperInputis_keras_tensor ZeroPadding2Dr3r1r2 MaxPooling2Drr9GlobalAveragePooling2Dvalidate_activationDenseGlobalMaxPooling2Drget_source_inputsrget_fileDENSENET121_WEIGHT_PATHDENSENET169_WEIGHT_PATHDENSENET201_WEIGHT_PATHDENSENET121_WEIGHT_PATH_NO_TOPDENSENET169_WEIGHT_PATH_NO_TOPDENSENET201_WEIGHT_PATH_NO_TOP load_weights)r include_toprQ input_tensor input_shapepoolingclassesclassifier_activationr img_inputr8rinputsmodel weights_paths rDenseNetrks3\  "&66 1  ) )Z->->w-G <  *D 1  !33--/# KLL{3 &&|4 L LI$I,,./AaqG6%56yAAJ b!Q\J1MA !! hZ  A 5&|4Q7A6%56q9A7Aqw7:AAvayw/ACg.AAvayw/ACg.AAvayw/ACg.AAvayw/AL!!wtLQOA.&v.q1A :F ) )z :1 =**+@'J FLL  5M   e >--:>qAA  :))z:1=A 22<@ vqt ,E* ()22G+!)@   ?*)22G+!)@   ?*)22G+!)@   !!9::()22M2!)@   ?*)22M2!)@   ?*)22M2!)@   !!9:: <( L   7# Lrz'keras.applications.densenet.DenseNet121zkeras.applications.DenseNet121c .tgd|||||||| S)z*Instantiates the Densenet121 architecture.rirrrrQrrrrrrs r DenseNet121rE,"    rz'keras.applications.densenet.DenseNet169zkeras.applications.DenseNet169c .tgd|||||||| S)z*Instantiates the Densenet169 architecture.rrrrrs r DenseNet169rcrrz'keras.applications.densenet.DenseNet201zkeras.applications.DenseNet201c .tgd|||||||| S)z*Instantiates the Densenet201 architecture.rtrrrs r DenseNet201rrrz,keras.applications.densenet.preprocess_inputc2tj||dS)Ntorch)rOmode)rpreprocess_input)rrOs rrrs  * * { rz.keras.applications.densenet.decode_predictionsc0tj||S)N)top)rdecode_predictions)predsrs rrrs  , ,U <rsK-1')&MKK?@ KK?@ KK?@   4 @   # Wt1(   #   01(   #   01(   #   0<=> >?=@=*>>EE 55  3 3F ,>>FF6p Y 3 3c 9: Y 3 3c 9: Y 3 3c 9:r