0h?dZddlmcmZddlmZddlmZddl m Z ddl m Z ddl mZddl mZdd lmZd Zd Ze Zed d  ddZ ddZedddZedddZej2j5dej6ej8e_ej0je_y)zInception V3 model for TF-Keras. Reference: - [Rethinking the Inception Architecture for Computer Vision]( http://arxiv.org/abs/1512.00567) (CVPR 2016) N)backend)imagenet_utils)training)VersionAwareLayers) data_utils) layer_utils) keras_exportz|https://storage.googleapis.com/tensorflow/keras-applications/inception_v3/inception_v3_weights_tf_dim_ordering_tf_kernels.h5zhttps://storage.googleapis.com/tensorflow/keras-applications/inception_v3/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5z+keras.applications.inception_v3.InceptionV3zkeras.applications.InceptionV3c |dvs7tjjj|st d||dk(r|r|dk7rt d|t j |ddtj||}|tj| }n/tj|stj|| }n|}tjd k(rd }nd }t|dd d dd} t| dd d d} t| dd d } tjdd| } t| dd d d} t| dd d d} tjdd| } t| dd d } t| dd d } t| ddd} t| dd d } t| dd d } t| dd d } tjddd| } t| dd d } tj| | | | g|d} t| dd d } t| dd d } t| ddd} t| dd d } t| dd d } t| dd d } tjddd| } t| dd d } tj| | | | g|d} t| dd d } t| dd d } t| ddd} t| dd d } t| dd d } t| dd d } tjddd| } t| dd d } tj| | | | g|d } t| d!d d dd}t| dd d } t| dd d } t| dd d dd} tjdd| } tj|| | g|d"} t| dd d } t| d#d d }t|d#d d$}t|dd$d }t| d#d d }t|d#d$d }t|d#d d$}t|d#d$d }t|dd d$}tjddd| } t| dd d } tj| ||| g|d%} t!d&D]}t| dd d } t| d'd d }t|d'd d$}t|dd$d }t| d'd d }t|d'd$d }t|d'd d$}t|d'd$d }t|dd d$}tjddd| } t| dd d } tj| ||| g|d(t#d|zz} t| dd d } t| dd d }t|dd d$}t|dd$d }t| dd d }t|dd$d }t|dd d$}t|dd$d }t|dd d$}tjddd| } t| dd d } tj| ||| g|d)} t| dd d }t|d*d d dd}t| dd d }t|dd d$}t|dd$d }t|dd d dd}tjdd| } tj||| g|d+} t!d&D]}t| d*d d } t| d!d d }t|d!d d }t|d!d d }tj||g|d,t#|z}t| d-d d } t| d!d d } t| d!d d }t| d!d d }tj||g|.} tjddd| } t| dd d } tj| || | g|d(t#d/|zz}  |rOtj%d01| } t j&||tj)||d23| } n=|d4k(rtj%| } n|d5k(rtj+| } |t-j.|}n|}t1j2|| d61}|dk(rP|rt5j6d7t8d8d9:}nt5j6d;t:d8d<:}|j=||S||j=||S)=a Instantiates the Inception v3 architecture. Reference: - [Rethinking the Inception Architecture for Computer Vision]( http://arxiv.org/abs/1512.00567) (CVPR 2016) This function returns a TF-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 TF-Keras Application expects a specific kind of input preprocessing. For `InceptionV3`, call `tf.keras.applications.inception_v3.preprocess_input` on your inputs before passing them to the model. `inception_v3.preprocess_input` will scale input pixels between -1 and 1. Args: include_top: Boolean, whether to include the fully-connected layer at the top, as the last layer of the network. Defaults to `True`. weights: One of `None` (random initialization), `imagenet` (pre-training on ImageNet), or the path to the weights file to be loaded. Defaults to `imagenet`. input_tensor: Optional TF-Keras tensor (i.e. output of `layers.Input()`) to use as image input for the model. `input_tensor` is useful for sharing inputs between multiple different networks. Defaults to `None`. input_shape: Optional shape tuple, only to be specified if `include_top` is False (otherwise the input shape has to be `(299, 299, 3)` (with `channels_last` data format) or `(3, 299, 299)` (with `channels_first` data format). It should have exactly 3 inputs channels, and width and height should be no smaller than 75. E.g. `(150, 150, 3)` would be one valid value. `input_shape` will be ignored if the `input_tensor` is provided. pooling: Optional pooling mode for feature extraction when `include_top` is `False`. - `None` (default) 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"`. Returns: A `keras.Model` instance. >NimagenetzThe `weights` argument should be either `None` (random initialization), `imagenet` (pre-training on ImageNet), or the path to the weights file to be loaded; Received: weights=r zjIf using `weights` as `"imagenet"` with `include_top` as true, `classes` should be 1000; Received classes=i+K) default_sizemin_size data_formatrequire_flattenweights)shape)tensorrchannels_first )rvalid)stridespadding)r@)rr)rP0`rrsamemixed0)axisnamemixed1mixed2imixed3mixed4rmixedmixed7i@mixed8mixed9_i)r& avg_poolr' predictions) activationr'avgmax inception_v3z2inception_v3_weights_tf_dim_ordering_tf_kernels.h5models 9a0d58056eeedaa3f26cb7ebd46da564) cache_subdir file_hashz8inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5 bcbd6486424b2319ff4ef7d526e38f63)tfiogfileexists ValueErrorrobtain_input_shaperimage_data_formatlayersInputis_keras_tensor conv2d_bn MaxPooling2DAveragePooling2D concatenaterangestrGlobalAveragePooling2Dvalidate_activationDenseGlobalMaxPooling2Drget_source_inputsrModelrget_file WEIGHTS_PATHWEIGHTS_PATH_NO_TOP load_weights) include_topr input_tensor input_shapepoolingclassesclassifier_activation img_input channel_axisx branch1x1 branch5x5 branch3x3dbl branch_pool branch3x3 branch7x7 branch7x7dbli branch7x7x3 branch3x3_1 branch3x3_2branch3x3dbl_1branch3x3dbl_2inputsmodel weights_paths `/var/www/html/engine/venv/lib/python3.12/site-packages/tf_keras/src/applications/inception_v3.py InceptionV3rt/sd T ) )RUU[[-?-?-H !")  +  *D 'y *  !33--/# KLL{L3 &&|4 L LI$I  "&66  )RAvwGA!RAw/A!RAAFF3A6A!RAw/A!S!Q0AFF3A6A!RA&I!RA&I)RA.IQAq)L\2q!4L\2q!4L))* KKQ2K I|[9    A!RA&I!RA&I)RA.IQAq)L\2q!4L\2q!4L))* KKQ2K I|[9    A!RA&I!RA&I)RA.IQAq)L\2q!4L\2q!4L))* KKQ2K I|[9    A!S!QHIQAq)L\2q!4Lb!QL%%ff%=a@K L+.\  A !S!Q'I!S!Q'I)S!Q/I)S!Q/IQQ*L\315L\315L\315L\315L))* KKa3K I|[9    A1X aa+ aa+ ia3 ia3  CA.  sAq9  sAq9  sAq9  sAq9 -- FF.   S!Q7     < =3q1u:%  # 0!S!Q'I!S!Q'I)S!Q/I)S!Q/IQQ*L\315L\315L\315L\315L))* KKa3K I|[9    A!S!Q'I)S!QPIAsAq)KKa3KKa3KS!QK%%ff%=a@K K-Lx  A 1X aa+ aa+  315  315 && + &SV#' !CA.  sAq9 "<a;"<a;)) ^ ,<* -- FF.   S!Q7     < =3q1u:%  1 :  ) )z ) :1 =**+@'J LL  5M    e --/2A  ))+A.A..|< NN61> :E* %..D%< L&..J#%< L <( L   7# Lc | |dz}|dz}nd}d}tjdk(rd} nd} tj|||f||d||}tj | d| |}tj d | |}|S) aUtility function to apply conv + BN. Args: x: input tensor. filters: filters in `Conv2D`. num_row: height of the convolution kernel. num_col: width of the convolution kernel. padding: padding mode in `Conv2D`. strides: strides in `Conv2D`. name: name of the ops; will become `name + '_conv'` for the convolution and `name + '_bn'` for the batch norm layer. Returns: Output tensor after applying `Conv2D` and `BatchNormalization`. N_bn_convrrrF)rruse_biasr')r&scaler'relur5)rrFrGConv2DBatchNormalization Activation) rbfiltersnum_rownum_colrrr'bn_name conv_namebn_axiss rsrJrJs& ,7N    "&66  '    A !!we'!J1MA&t,Q/A Hruz0keras.applications.inception_v3.preprocess_inputc2tj||dS)Nr@)rmode)rpreprocess_input)rbrs rsrrs  * * { ruz2keras.applications.inception_v3.decode_predictionsc0tj||S)N)top)rdecode_predictions)predsrs rsrrs  , ,U <rs "! 4(2)*:F L  1$    #^ ^D HL' T@AB BC=D=*>>EE 22  3 3F ,>>FFru