~WhyBdZddlmcmZddlmZddlmZddl m Z ddl m Z ddl mZddl mZdd lmZdd lmZd Zed zZed zZedzZedzZe Z d,dZedd d-dZedd d-dZ d.d!Zd/d"Zd/d#Z d/d$Z!ed%d/d&Z"ed'd0d)Z#ej$%d*ej&ej'+e"_ej#je#_dS)1aGNASNet-A models for Keras. NASNet refers to Neural Architecture Search Network, a family of models that were designed automatically by learning the model architectures directly on the dataset of interest. Here we consider NASNet-A, the highest performance model that was found for the CIFAR-10 dataset, and then extended to ImageNet 2012 dataset, obtaining state of the art performance on CIFAR-10 and ImageNet 2012. Only the NASNet-A models, and their respective weights, which are suited for ImageNet 2012 are provided. The below table describes the performance on ImageNet 2012: --------------------------------------------------------------------------- Architecture | Top-1 Acc | Top-5 Acc | Multiply-Adds | Params (M) ---------------------|-----------|-----------|----------------|------------ NASNet-A (4 @ 1056) | 74.0 % | 91.6 % | 564 M | 5.3 NASNet-A (6 @ 4032) | 82.7 % | 96.2 % | 23.8 B | 88.9 Reference: - [Learning Transferable Architectures for Scalable Image Recognition]( https://arxiv.org/abs/1707.07012) (CVPR 2018) N)backend)imagenet_utils)training)VersionAwareLayers) data_utils) layer_utils) tf_logging) keras_exportzDhttps://storage.googleapis.com/tensorflow/keras-applications/nasnet/zNASNet-mobile.h5zNASNet-mobile-no-top.h5zNASNet-large.h5zNASNet-large-no-top.h5`Timagenetsoftmaxc  |dvs3tjj|st d|dkr|r| dkrt dt |t r,d|vr(|dkr"t dt|zdz| d } tj || d tj || }tj d kr+tj d tjd d} nd} |t|}n3tj|st||}n|}|d|dzzzdkrt d|ztj dkrdnd}|dz}t|dddddd|}t|ddd !|}d}t)||||dzzd"#\}}t)||||zd$#\}}t+|D]}t-|||d%|z#\}}t)||||zd&|z#\}}|s|n|}t+|D]$}t-||||zd%||zdzz#\}}%t)||||dzzd&d|zz#\}}|s|n|}t+|D]*}t-||||dzzd%d|z|zdzz#\}}+td'|}|r\t|}tj| |t| | d()|}nO| d*kr"t|}n'| d+kr!t|}|t9j|}n|}t=j||d,-}|dkr| d.krS|rtAj!d/tDd0d12}ntAj!d3tFd0d42}|$|n| d krS|rtAj!d5tJd0d62}ntAj!d7tLd0d82}|$|n&t d9||$|| rtj| |S):aInstantiates a NASNet model. Reference: - [Learning Transferable Architectures for Scalable Image Recognition]( https://arxiv.org/abs/1707.07012) (CVPR 2018) 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 NasNet, call `tf.keras.applications.nasnet.preprocess_input` on your inputs before passing them to the model. `nasnet.preprocess_input` will scale input pixels between -1 and 1. Args: input_shape: Optional shape tuple, the input shape is by default `(331, 331, 3)` for NASNetLarge and `(224, 224, 3)` for NASNetMobile. It should have exactly 3 input channels, and width and height should be no smaller than 32. E.g. `(224, 224, 3)` would be one valid value. penultimate_filters: Number of filters in the penultimate layer. NASNet models use the notation `NASNet (N @ P)`, where: - N is the number of blocks - P is the number of penultimate filters num_blocks: Number of repeated blocks of the NASNet model. NASNet models use the notation `NASNet (N @ P)`, where: - N is the number of blocks - P is the number of penultimate filters stem_block_filters: Number of filters in the initial stem block skip_reduction: Whether to skip the reduction step at the tail end of the network. filter_multiplier: Controls the width of the network. - If `filter_multiplier` < 1.0, proportionally decreases the number of filters in each layer. - If `filter_multiplier` > 1.0, proportionally increases the number of filters in each layer. - If `filter_multiplier` = 1, default number of filters from the paper are used at each layer. include_top: Whether to include the fully-connected layer at the top of the network. weights: `None` (random initialization) or `imagenet` (ImageNet weights) input_tensor: Optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model. 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. default_size: Specifies the default image size of the model 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. >NrzThe `weights` argument should be either `None` (random initialization), `imagenet` (pre-training on ImageNet), or the path to the weights file to be loaded.rrzWIf using `weights` as `"imagenet"` with `include_top` as true, `classes` should be 1000NzWhen specifying the input shape of a NASNet and loading `ImageNet` weights, the input_shape argument must be static (no None entries). Got: `input_shape=z`.K ) default_sizemin_size data_formatrequire_flattenweights channels_lastaThe NASNet family of models is only available for the input data format "channels_last" (width, height, channels). However your settings specify the default data format "channels_first" (channels, width, height). You should set `image_data_format="channels_last"` in your Keras config located at ~/.keras/keras.json. The model being returned right now will expect inputs to follow the "channels_last" data format.channels_first)shape)tensorrrrzwFor NASNet-A models, the `penultimate_filters` must be a multiple of 24 * (`filter_multiplier` ** 2). Current value: %dr"rrvalidF stem_conv1 he_normal)stridespaddinguse_biasnamekernel_initializerZڊ?MbP?stem_bn1axismomentumepsilonr*stem_1block_idstem_2z%dz reduce_%drelu predictions) activationr*avgmaxNASNetr*znasnet_mobile.h5models 020fb642bf7360b370c678b08e0adf61) cache_subdir file_hashznasnet_mobile_no_top.h5 1ed92395b5b598bdda52abe5c0dbfd63znasnet_large.h5 11577c9a518f0070763c2b964a382f17znasnet_large_no_top.h5 d81d89dc07e6e56530c4e77faddd61b5zDImageNet weights can only be loaded with NASNetLarge or NASNetMobile)'tfiogfileexists ValueError isinstancetuplestrrobtain_input_shaperimage_data_formatloggingwarningset_image_data_formatlayersInputis_keras_tensorConv2DBatchNormalization_reduction_a_cellrange_normal_a_cell ActivationGlobalAveragePooling2Dvalidate_activationDenseGlobalMaxPooling2Drget_source_inputsrModelrget_fileNASNET_MOBILE_WEIGHT_PATH NASNET_MOBILE_WEIGHT_PATH_NO_TOP load_weightsNASNET_LARGE_WEIGHT_PATHNASNET_LARGE_WEIGHT_PATH_NO_TOP) input_shapepenultimate_filters num_blocksstem_block_filtersskip_reductionfilter_multiplier include_topr input_tensorpoolingclassesrclassifier_activationold_data_format img_input channel_dimfiltersxpip0inputsmodel weights_paths _/var/www/html/movieo_spanner_bot/venv/lib/python3.11/site-packages/keras/applications/nasnet.pyr<r<@sv ) ) )RU[-?-?-H-H ) <   *D 0   ;&&  K   z ! ! 47:+6F6F GJN N    !3!-//# K ""o55 9  %o666*LL{L33 &|44 % L LLII$Ib$5q$89:a?? D! "   0226FFF!!BK!R'G  &     A !! 64j "    A A  1g+Q./(   DAq  1g**X   DAq:  BBaGdajAAA11  1g))K:4N   EAr!'aA :     ' 'Z!^a/0    11   #Q&&J/    EAr!'aA :     '* *Q^a/!34    11 &!!!$$A /  ) ) + +A . .*+@'JJJ LL  5M      e  --//22AA   ))++A..A.|<< N618 4 4 4E* 3   )2&-!)@      *2-4!)@       | , , , , S  )2%,!)@      *2,3!)@       | , , , ,#    7###7%o666 Lz&keras.applications.nasnet.NASNetMobilezkeras.applications.NASNetMobilec:t|ddddd|||||d| S)a Instantiates a Mobile NASNet model in ImageNet mode. Reference: - [Learning Transferable Architectures for Scalable Image Recognition]( https://arxiv.org/abs/1707.07012) (CVPR 2018) Optionally loads weights pre-trained on ImageNet. Note that the data format convention used by the model is the one specified in your Keras config at `~/.keras/keras.json`. Note: each Keras Application expects a specific kind of input preprocessing. For NASNet, call `tf.keras.applications.nasnet.preprocess_input` on your inputs before passing them to the model. Args: input_shape: Optional shape tuple, only to be specified if `include_top` is False (otherwise the input shape has to be `(224, 224, 3)` for NASNetMobile It should have exactly 3 inputs channels, and width and height should be no smaller than 32. E.g. `(224, 224, 3)` would be one valid value. include_top: Whether to include the fully-connected layer at the top of the network. weights: `None` (random initialization) or `imagenet` (ImageNet weights). For loading `imagenet` weights, `input_shape` should be (224, 224, 3) input_tensor: Optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model. 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"`. Returns: A Keras model instance. Raises: ValueError: In case of invalid argument for `weights`, or invalid input shape. RuntimeError: If attempting to run this model with a backend that does not support separable convolutions. i rFrr> rirjrkrlrmrnrrorprqrrrr<rhrnrrorprqrrs r~ NASNetMobilerdsCN  !3   rz%keras.applications.nasnet.NASNetLargezkeras.applications.NASNetLargec:t|ddddd|||||d| S)a Instantiates a NASNet model in ImageNet mode. Reference: - [Learning Transferable Architectures for Scalable Image Recognition]( https://arxiv.org/abs/1707.07012) (CVPR 2018) Optionally loads weights pre-trained on ImageNet. Note that the data format convention used by the model is the one specified in your Keras config at `~/.keras/keras.json`. Note: each Keras Application expects a specific kind of input preprocessing. For NASNet, call `tf.keras.applications.nasnet.preprocess_input` on your inputs before passing them to the model. Args: input_shape: Optional shape tuple, only to be specified if `include_top` is False (otherwise the input shape has to be `(331, 331, 3)` for NASNetLarge. It should have exactly 3 inputs channels, and width and height should be no smaller than 32. E.g. `(224, 224, 3)` would be one valid value. include_top: Whether to include the fully-connected layer at the top of the network. weights: `None` (random initialization) or `imagenet` (ImageNet weights). For loading `imagenet` weights, `input_shape` should be (331, 331, 3) input_tensor: Optional Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model. 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"`. Returns: A Keras model instance. Raises: ValueError: in case of invalid argument for `weights`, or invalid input shape. RuntimeError: If attempting to run this model with a backend that does not support separable convolutions. r r r Trrrrrs r~ NASNetLargersCN  !3   rr!rrc tjdkrdnd}tjd|5td|}|dkr=tt j||d||}d }nd }t|||d ||d d |}t |ddd||}td|}t||d|d d d |}t |ddd||}dddn #1swxYwY|S)aFAdds 2 blocks of [relu-separable conv-batchnorm]. Args: ip: Input tensor filters: Number of output filters per layer kernel_size: Kernel size of separable convolutions strides: Strided convolution for downsampling block_id: String block_id Returns: A Keras tensor rrr separable_conv_block_r7r#separable_conv_1_pad_r(r*r$sameseparable_conv_1_Fr&)r'r*r(r)r+r,r-separable_conv_1_bn_r/separable_conv_2_)r*r(r)r+separable_conv_2_bn_N) rrO name_scoperSr[ ZeroPadding2Dr correct_padSeparableConv2DrW)iprv kernel_sizer'r5rurwconv_pads r~_separable_conv_blockrs50226FFF!!BK  >H>> ? ?' '   f % %b ) ) f  $$&21kBB7X77%AHHH  " "  /X//* #       % %222 &       f % %a ( (  " "  /X//* #       % %222 &    E' ' ' ' ' ' ' ' ' ' ' ' ' ' ' P HsD3E22E69E6c tjdkrdnd}tjdkrdnd}tj|}|tj|}tjd5||}nW||||krtjd|5t d d | |}t d d dd||}t |dzd ddd|d|}t d|} t d| } t d d dd|| } t |dzd ddd|d| } t || g|}t |ddd| |}dddn #1swxYwYn|||krtjd!|5t d |}t |d d dd"|dd#|}t |ddd| |}dddn #1swxYwYdddn #1swxYwY|S)$aAdjusts the input `previous path` to match the shape of the `input`. Used in situations where the output number of filters needs to be changed. Args: p: Input tensor which needs to be modified ip: Input tensor whose shape needs to be matched filters: Number of output filters to be matched block_id: String block_id Returns: Adjusted Keras tensor rrr rN adjust_blockadjust_reduction_block_r7adjust_relu_1_r=rr#r$adjust_avg_pool_1_r'r(r*rFadjust_conv_1_r&)r(r)r*r+)rrr)r()rrr)croppingadjust_avg_pool_2_adjust_conv_2_)r0r,r- adjust_bn_r/adjust_projection_block_adjust_conv_projection_r'r(r*r)r+) rrO int_shaperrSr[AveragePooling2DrVr Cropping2D concatenaterW) rxrrvr5ruimg_dimip_shapep_shapep1p2s r~ _adjust_blockrPsh0226FFF!!BK,..2BBBaaG $$H}#A&&  N + +BB 9AA W '!2 2 2#$Hh$H$HII* * %%f3NH3N3N%OO,,"#8h88 -   ]]qL""4(44'2 #))2B)CCAFF&&0@&AA"EE,,"#8h88 -   ]]qL""4(44'2 #&&Bxk&BB--$# 0h00 .  K* * * * * * * * * * * * * * * X[ !W , ,#$Ix$I$IJJ  %%f--a00MM""=8=="'2"--$# 0h00 .                 eBBBBBBBBBBBBBBBF Hs\51K&E#H KH KH 'KA8K< KK KK KK#&K#c tjdkrdnd}tjd|5t||||}t d|}t |dddd|d d |}t |d d d||}tjd5t||dd|}t||d|}t ||gd|}dddn #1swxYwYtjd5t||dd|} t||dd|} t | | gd|} dddn #1swxYwYtjd5t dddd| |} t | |gd!|} dddn #1swxYwYtjd"5t dddd#| |} t dddd$| |}t | |gd%|}dddn #1swxYwYtjd&5t||d'|}t ||gd(|}dddn #1swxYwYt ||| | ||g|d)|*}dddn #1swxYwY||fS)+zAdds a Normal cell for NASNet-A (Fig. 4 in the paper). Args: ip: Input tensor `x` p: Input tensor `p` filters: Number of output filters block_id: String block_id Returns: A Keras tensor rrr normal_A_block_r7rrnormal_conv_1_Fr&rr,r- normal_bn_1_r/block_1r normal_left1_)rr5normal_right1_r4 normal_add_1_r=Nblock_2 normal_left2_r!normal_right2_ normal_add_2_block_3 normal_left3_r normal_add_3_block_4 normal_left4_normal_right4_ normal_add_4_block_5 normal_left5_ normal_add_5_normal_concat_r0r*) rrOrrrSr[rVrWraddrr)rrxrvr5ruhx1_1x1_2x1x2_1x2_2x2x3x4_1x4_2x4x5rws r~rZrZs#0226FFF!!BK  8h88 9 9K K !R( 3 3   f % %b ) ) MM  ,(,,*        % %*** &      * * K K("333 D )7%@h%@%@DT4L/Ix/I/IJJB K K K K K K K K K K K K K K K  * * K K(7F-GX-G-GD)7F-Hh-H-HDT4L/Ix/I/IJJB K K K K K K K K K K K K K K K  * * F F((/X// )  B RG*D(*D*DEEB F F F F F F F F F F F F F F F  * * K K**/X// +  D **0h00 +  D T4L/Ix/I/IJJB K K K K K K K K K K K K K K K  * * F F&7%?X%?%?BRG*D(*D*DEEB  F F F F F F F F F F F F F F F    BB #,(,,   OK K K K K K K K K K K K K K K X b5LsBMA D* M*D. .M1D. 2M AF# M#F' 'M*F' +MA H MH M H !M8A4J8, M8J< <M?J< M7L ML M!L ")MMMc tjdkrdnd}tjd|5t||||}t d|}t |dddd|d d |}t |d d d||}t tj |dd||}tjd5t||ddd|}t||ddd|}t ||gd|} dddn #1swxYwYtjd5t dddd |!|} t||ddd"|} t | | gd#|} dddn #1swxYwYtjd$5t dddd%|!|} t||ddd&|}t | |gd'|}dddn #1swxYwYtjd(5t dddd)|!| }t | |g}dddn #1swxYwYtjd*5t| |dd)|+}t dddd,|!|}t ||gd-|}dddn #1swxYwYt| |||g|d.|/}||fcdddS#1swxYwYdS)0zAdds a Reduction cell for NASNet-A (Fig. 4 in the paper). Args: ip: Input tensor `x` p: Input tensor `p` filters: Number of output filters block_id: String block_id Returns: A Keras tensor rrr reduction_A_block_r7rrreduction_conv_1_Fr&rr,r-reduction_bn_1_r/r"reduction_pad_1_rrrr#reduction_left1_)r'r5)rreduction_right1_reduction_add_1_r=Nrr!r$reduction_left2_rreduction_right2_reduction_add_2_rreduction_left3_reduction_right3_reduction_add3_rreduction_left4_rr4reduction_right5_reduction_add4_reduction_concat_r)rrOrrrSr[rVrWrrrrr MaxPooling2Drr)rrxrvr5rurh3rrrrrrx3_1x3_2rrx5_1x5_2rrws r~rXrX s0226FFF!!BK  ;;; < <cc !R( 3 3   f % %b ) ) MM  /X//*        % %-8-- &      ! !".q!44.H.."        * * N N(6H66 D)7X77 DT4L/L(/L/LMMB N N N N N N N N N N N N N N N"  * * N N&&222 '  D )7X77 DT4L/L(/L/LMMB N N N N N N N N N N N N N N N   * * M M**222 +  D )7X77 DT4L/K/K/KLLB M M M M M M M M M M M M M M M   * * & &((222 )  B RH%%B & & & & & & & & & & & & & & &  * * M M(GV.K.K.KD&&333 '  D T4L/K/K/KLLB M M M M M M M M M M M M M M M    R /X//    "uGccccccccccccccccccsCN= AE' N='E+ +N=.E+ /N=A"G4( N=4G8 8N=;G8 <N=A"J5 N=J N=J N= AK2& N=2K6 6N=9K6 :N=A!M>2 N=>N N=N *N==OOz*keras.applications.nasnet.preprocess_inputc0tj||dS)NrF)rmode)rpreprocess_input)rwrs r~rr}s#  * {   rz,keras.applications.nasnet.decode_predictionsrc.tj||S)N)top)rdecode_predictions)predsrs r~rrs  ,U < < <rs 0"!!!!!!!!------!!!!!!++++++""""""######=<<<<<999999K.0BB#47P#P ,/@@"36N"N       #aaaaH ,.O   #RRRRj+-M   #RRRRl?C9 9 9 9 xY Y Y Y xZZZZzqqqqh:;;<; <=====>==*>EE 2  3F ,>Fr