&VfjjddlmZddlmZddlmZddlmZddlmZddlmZddlmZddlm Z dd l m Z dd l m Z dd lmZdd lmZdd lmZGdde e ZGddeZdS)) activations)backend) constraints) initializers)ops) regularizers)tree) InputSpec)Layer)DropoutRNNCell)RNN)operation_utils)argument_validationceZdZdZ dfd Zdd ZddZddZddZddZ dZ fdZ xZ S) ConvLSTMCellaCell class for the ConvLSTM layer. Args: rank: Integer, rank of the convolution, e.g. "2" for 2D convolutions. filters: Integer, the dimensionality of the output space (i.e. the number of output filters in the convolution). kernel_size: An integer or tuple/list of n integers, specifying the dimensions of the convolution window. strides: An integer or tuple/list of n integers, specifying the strides of the convolution. Specifying any stride value != 1 is incompatible with specifying any `dilation_rate` value != 1. padding: One of `"valid"` or `"same"` (case-insensitive). `"valid"` means no padding. `"same"` results in padding evenly to the left/right or up/down of the input such that output has the same height/width dimension as the input. data_format: A string, one of `channels_last` (default) or `channels_first`. When unspecified, uses `image_data_format` value found in your Keras config file at `~/.keras/keras.json` (if exists) else 'channels_last'. Defaults to `'channels_last'`. dilation_rate: An integer or tuple/list of n integers, specifying the dilation rate to use for dilated convolution. Currently, specifying any `dilation_rate` value != 1 is incompatible with specifying any `strides` value != 1. activation: Activation function. If `None`, no activation is applied. recurrent_activation: Activation function to use for the recurrent step. use_bias: Boolean, (default `True`), whether the layer should use a bias vector. kernel_initializer: Initializer for the `kernel` weights matrix, used for the linear transformation of the inputs. Default: `"glorot_uniform"`. recurrent_initializer: Initializer for the `recurrent_kernel` weights matrix, used for the linear transformation of the recurrent state. Default: `"orthogonal"`. bias_initializer: Initializer for the bias vector. Default: `"zeros"`. unit_forget_bias: Boolean (default `True`). If `True`, add 1 to the bias of the forget gate at initialization. Setting it to `True` will also force `bias_initializer="zeros"`. This is recommended in [Jozefowicz et al.]( https://github.com/mlresearch/v37/blob/gh-pages/jozefowicz15.pdf) kernel_regularizer: Regularizer function applied to the `kernel` weights matrix. Default: `None`. recurrent_regularizer: Regularizer function applied to the `recurrent_kernel` weights matrix. Default: `None`. bias_regularizer: Regularizer function applied to the bias vector. Default: `None`. activity_regularizer: Regularizer function applied to the output of the layer (its "activation"). Default: `None`. kernel_constraint: Constraint function applied to the `kernel` weights matrix. Default: `None`. recurrent_constraint: Constraint function applied to the `recurrent_kernel` weights matrix. Default: `None`. bias_constraint: Constraint function applied to the bias vector. Default: `None`. dropout: Float between 0 and 1. Fraction of the units to drop for the linear transformation of the inputs. Default: 0. recurrent_dropout: Float between 0 and 1. Fraction of the units to drop for the linear transformation of the recurrent state. Default: 0. seed: Random seed for dropout. Call arguments: inputs: A (2+ `rank`)D tensor. states: List of state tensors corresponding to the previous timestep. training: Python boolean indicating whether the layer should behave in training mode or in inference mode. Only relevant when `dropout` or `recurrent_dropout` is used. validNtanhsigmoidTglorot_uniform orthogonalzerosc "tjdi|||_tj||_||_|jdkrtd|d|||_ tj ||jd|_ tj ||jdd|_ tj||_tj||_tj ||jd |_t'j||_t'j| |_| |_t1j| |_t1j| |_t1j| |_||_t;j||_t;j||_t;j||_ tCj||_"tCj||_#tCj||_$tKd tMd ||_'tKd tMd ||_(tS|d z |_*d|_+dS)N)seedzRank z< convolutions are not currently implemented. Received: rank= kernel_sizestridesT allow_zero dilation_rateg?rndim),super__init__rrrandom SeedGeneratorseed_generatorrank ValueErrorfiltersrstandardize_tuplerrstandardize_paddingpaddingstandardize_data_format data_formatr!rget activationrecurrent_activationuse_biasrkernel_initializerrecurrent_initializerbias_initializerunit_forget_biasrkernel_regularizerrecurrent_regularizerbias_regularizerrkernel_constraintrecurrent_constraintbias_constraintminmaxdropoutrecurrent_dropoutr input_spec state_size)selfr,r.rrr1r3r!r5r6r7r8r9r:r;r<r=r>r?r@rArDrErkwargs __class__s [/var/www/html/software/conda/lib/python3.11/site-packages/keras/src/layers/rnn/conv_lstm.pyr(zConvLSTMCell.__init__UsV6 ""6""" %n:::EE 9q==666/366  .@ M  +< TY d    +>wGG ":;GG0B 49o  &/*55$/O4H$I$I!  "."23E"F"F%1%56K%L%L" , 01A B B 0"."23E"F"F%1%56K%L%L" , 01A B B!,1B!C!C$/O4H$I$I!*??3C 1 122 !$S#c3D*E*E!F!F#222cjdkrd}|dd_nd}|dd_djvrtd|||td|tjdzd|ddz _||}|_j|jd zfz_ jjjd zfz} j j d j j _ |jd jj _jrGjrfd}nj} jd zfd|jj_nd_d_dS)Nchannels_firstrr"r%zvConvLSTM layers only support static input shapes for the spatial dimension. Received invalid input shape: input_shape=zvThe channel dimension of the inputs (last axis) should be defined. Found None. Full input shape received: input_shape=rN)r$shapekernel)rP initializername regularizer constraintrecurrent_kernelc tjjjfg|Ri|t jdjfg|Ri|jjdzfg|Ri|gS)Nonesr")r concatenater:r.rr4)_argsrIrHs rKr:z,ConvLSTMCell.build..bias_initializers?1D1!%26:@5L,V44!%26:@2D1!%!1 36:>D    rLbias)rPrTrSrUrVT)r3 spatial_dimsr-r r,rF input_dimrr. kernel_shape add_weightr8r<r?rRr9r=r@rWr7r;r:r>rAr]built)rH inputs_shape states_shape channel_axisr_recurrent_kernel_shaper:s` rKbuildzConvLSTMCell.builds$  / / /L ,QRR 0D  L ,QrT 2D  4$ $ $L=ILL   % -.+..  $Qg QRR0@&@   !. " , 4>!!'+=>>!!'+=>>!!'+=>>  % %cCi 0 0  % %cCi 0 0 IDOOC#I666 6  % %cCi 0 0 "" "1a&yrLc tj||j|j|j|j|j|j}|||gfS)Nrr1r3r!)rcompute_conv_output_shaper.rrr1r3r!)rHrcrdconv_output_shapes rKcompute_output_shapez!ConvLSTMCell.compute_output_shapesR+E  L  LL(,   !#46G"HHHrLc|jdkr|f|jz|jfz}n||jf|jz}||d}t j||jt j||jgS)N channels_lastr)dtype)r3r^r_rrr compute_dtype)rH batch_size input_shape state_shapes rKget_initial_statezConvLSTMCell.get_initial_state*s   . .%-$*;;t~>OOKK%t~69JJK// <r?)r@rArDrEr)r.rrr1r3r!r serializer5r6r7rr8r9r:r;rr<r=r>rr?r@rArDrErr' get_configrHconfig base_configrJs rKrzConvLSTMCell.get_configPs# t|# 4+#  t|#  t| # 4+ # T/ #  +/@@#  #K$9)%%#   #  !,"8'###  $\%;*&&# $  6t7L M M%# &  5'# ( !,"8'##)# . $\%;*&&/# 4  6t7L M M5# 6 !6t7M!N!N7# 8%0$9)%% +4T5IJJ|!%!7IE# # # Hgg((** (+(((rL)rrNrrrTrrrTNNNNNNrrNrO)F)Nr) __name__ __module__ __qualname____doc__r(rgrrrrkrlr __classcell__rJs@rKrrs%BBR&+* "! 1DDDDDDLKKKKZ4444l I I I I     $&)&)&)&)&)&)&)&)&)rLrc\eZdZdZ d(fd Zd)fd Zd*dZdZedZ edZ edZ edZ edZ edZedZedZedZedZedZedZedZedZedZed Zed!Zed"Zed#Zed$Zed%Zfd&Zed'Z xZ!S)+ConvLSTMaSAbstract N-D Convolutional LSTM layer (used as implementation base). Similar to an LSTM layer, but the input transformations and recurrent transformations are both convolutional. Args: rank: Integer, rank of the convolution, e.g. "2" for 2D convolutions. filters: Integer, the dimensionality of the output space (i.e. the number of output filters in the convolution). kernel_size: An integer or tuple/list of n integers, specifying the dimensions of the convolution window. strides: An integer or tuple/list of n integers, specifying the strides of the convolution. Specifying any stride value != 1 is incompatible with specifying any `dilation_rate` value != 1. padding: One of `"valid"` or `"same"` (case-insensitive). `"valid"` means no padding. `"same"` results in padding evenly to the left/right or up/down of the input such that output has the same height/width dimension as the input. data_format: A string, one of `channels_last` (default) or `channels_first`. The ordering of the dimensions in the inputs. `channels_last` corresponds to inputs with shape `(batch, time, ..., channels)` while `channels_first` corresponds to inputs with shape `(batch, time, channels, ...)`. When unspecified, uses `image_data_format` value found in your Keras config file at `~/.keras/keras.json` (if exists) else 'channels_last'. Defaults to `'channels_last'`. dilation_rate: An integer or tuple/list of n integers, specifying the dilation rate to use for dilated convolution. Currently, specifying any `dilation_rate` value != 1 is incompatible with specifying any `strides` value != 1. activation: Activation function to use. By default hyperbolic tangent activation function is applied (`tanh(x)`). recurrent_activation: Activation function to use for the recurrent step. use_bias: Boolean, whether the layer uses a bias vector. kernel_initializer: Initializer for the `kernel` weights matrix, used for the linear transformation of the inputs. recurrent_initializer: Initializer for the `recurrent_kernel` weights matrix, used for the linear transformation of the recurrent state. bias_initializer: Initializer for the bias vector. unit_forget_bias: Boolean. If True, add 1 to the bias of the forget gate at initialization. Use in combination with `bias_initializer="zeros"`. This is recommended in [Jozefowicz et al., 2015]( http://www.jmlr.org/proceedings/papers/v37/jozefowicz15.pdf) kernel_regularizer: Regularizer function applied to the `kernel` weights matrix. recurrent_regularizer: Regularizer function applied to the `recurrent_kernel` weights matrix. bias_regularizer: Regularizer function applied to the bias vector. activity_regularizer: Regularizer function applied to. kernel_constraint: Constraint function applied to the `kernel` weights matrix. recurrent_constraint: Constraint function applied to the `recurrent_kernel` weights matrix. bias_constraint: Constraint function applied to the bias vector. dropout: Float between 0 and 1. Fraction of the units to drop for the linear transformation of the inputs. recurrent_dropout: Float between 0 and 1. Fraction of the units to drop for the linear transformation of the recurrent state. seed: Random seed for dropout. return_sequences: Boolean. Whether to return the last output in the output sequence, or the full sequence. (default False) return_state: Boolean Whether to return the last state in addition to the output. (default False) go_backwards: Boolean (default False). If True, process the input sequence backwards. stateful: Boolean (default False). If True, the last state for each sample at index i in a batch will be used as initial state for the sample of index i in the following batch. rrNrrTrrrrFc Ftdid|d|d|d|d|d|d|d|d | d | d | d | d | d|d|d|d|d|d|d|d|d|d|ddd|d}tj|f||||d|t |dz|_dS)Nr,r.rrr1r3r!r5r6r7r8r9r:r;r<r=r>r?r@rArDrErrTconv_lstm_cellr)return_sequences return_state go_backwardsstatefulrr#r&)rr4r'r(r rF)rHr,r.rrr1r3r!r5r6r7r8r9r:r;r<r=r>r?r@rArDrErrrrrrIcellrJs rKr(zConvLSTM.__init__s>    G $  G  G  $  (- "z "6!5 X  21 #8"7 .- .-  21 #8"7! ".-# $0/% &"6!5' (,O) *G+ ,0/- ./ 0"!1 2**W%%%3 6   -%%       $222rLcNt||||S)N) initial_statemaskro)r'r)rH sequencesrrrorJs rKrz ConvLSTM.calls+ww|| ]   rLc|d}|d}|f|ddz}|j|ddd}|jr||f|z}n|f|z}|jr |f|z}|||fS|S)Nrrr")rrrr) rHsequences_shapeinitial_state_shapersteps step_shaper output_shapebatched_state_shapes rKrzConvLSTM.compute_output_shapes$Q' " ]_QRR%88 i44Z@@CABBG   7LL '=;6L   J#--+"= !46II IrLcrtj|d}|jr|nd}|jr ddg}|g|zS|S)Nr)r flattenrr)rHr[r output_mask state_masks rK compute_maskzConvLSTM.compute_mask&sN|D!!!$"3=dd   J=:- - rLc|jjSrO)rr.rHs rKr.zConvLSTM.filters/ y  rLc|jjSrO)rrrs rKrzConvLSTM.kernel_size3 y$$rLc|jjSrO)rrrs rKrzConvLSTM.strides7rrLc|jjSrO)rr1rs rKr1zConvLSTM.padding;rrLc|jjSrO)rr3rs rKr3zConvLSTM.data_format?rrLc|jjSrO)rr!rs rKr!zConvLSTM.dilation_rateCs y&&rLc|jjSrO)rr5rs rKr5zConvLSTM.activationGs y##rLc|jjSrO)rr6rs rKr6zConvLSTM.recurrent_activationK y--rLc|jjSrO)rr7rs rKr7zConvLSTM.use_biasOs y!!rLc|jjSrO)rr8rs rKr8zConvLSTM.kernel_initializerS y++rLc|jjSrO)rr9rs rKr9zConvLSTM.recurrent_initializerW y..rLc|jjSrO)rr:rs rKr:zConvLSTM.bias_initializer[ y))rLc|jjSrO)rr;rs rKr;zConvLSTM.unit_forget_bias_rrLc|jjSrO)rr<rs rKr<zConvLSTM.kernel_regularizercrrLc|jjSrO)rr=rs rKr=zConvLSTM.recurrent_regularizergrrLc|jjSrO)rr>rs rKr>zConvLSTM.bias_regularizerkrrLc|jjSrO)rr?rs rKr?zConvLSTM.kernel_constrainto y**rLc|jjSrO)rr@rs rKr@zConvLSTM.recurrent_constraintsrrLc|jjSrO)rrArs rKrAzConvLSTM.bias_constraintws y((rLc|jjSrO)rrDrs rKrDzConvLSTM.dropout{rrLc|jjSrO)rrErs rKrEzConvLSTM.recurrent_dropoutrrLcrid|jd|jd|jd|jd|jd|jdt j|jdt j|j d |j d tj|j d tj|j d tj|jd |jdt!j|jdt!j|jdt!j|jdt!j|jt+j|jt+j|jt+j|j|j|j|jjd}t;}|d=i||S)Nr.rrr1r3r!r5r6r7r8r9r:r;r<r=r>activity_regularizer)r?r@rArDrErr)r.rrr1r3r!rrr5r6r7rr8r9r:r;rr<r=r>rrr?r@rArDrErrr'rrs rKrzConvLSTM.get_configs& t|& 4+&  t|&  t| & 4+ & T/ &  +/@@&  #K$9)%%&   &  !,"8'##&  $\%;*&&& $  6t7L M M%& &  5'& ( !,"8'##)& . $\%;*&&/& 4  6t7L M M5& 6 #L$:)%%7& <"-!6t7M!N!N$/$9)%% +4T5IJJ|!%!7INK& & & Ngg((**  (+(((rLc|di|S)Nr&r&)clsrs rK from_configzConvLSTM.from_configss}}V}}rL)rrNrrrTrrrTNNNNNNrrNFFFF)NNFrO)"rrrrr(rrrpropertyr.rrr1r3r!r5r6r7r8r9r:r;r<r=r>r?r@rArDrEr classmethodrrrs@rKrrysHNNj&+* "! 9B3B3B3B3B3B3H      &!!X!%%X%!!X!!!X!%%X%''X'$$X$..X.""X",,X,//X/**X***X*,,X,//X/**X*++X+..X.))X)!!X!++X+*)*)*)*)*)X[rLrN) keras.srcrrrrrrr keras.src.layers.input_specr keras.src.layers.layerr %keras.src.layers.rnn.dropout_rnn_cellr keras.src.layers.rnn.rnnr keras.src.opsrkeras.src.utilsrrrr&rLrKrsb!!!!!!!!!!!!""""""""""""111111((((((@@@@@@(((((())))))//////f)f)f)f)f)5.f)f)f)R xxxxxsxxxxxrL