o 2hk@sddlmZddlmZddlmZddlmZddlmZddlmZddlmZddlm Z dd l m Z dd l m Z dd lmZdd lmZdd lmZGddde e ZGdddeZdS)) activations)backend) constraints) initializers)ops) regularizers)tree) InputSpec)Layer)DropoutRNNCell)RNN)operation_utils)argument_validationcseZdZdZ                dfd d ZdddZdddZdddZdddZd ddZ ddZ fddZ Z 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 sxtjdi|||_tjj|d|_||_|jdkr&td|d|||_ t ||jd|_ t j ||jddd|_ t ||_t||_t ||jd |_t||_t| |_| |_t| |_t| |_t| |_||_t||_t||_t||_ t!||_"t!||_#t!||_$t%d t&d ||_'t%d t&d ||_(d |_)t*|d d|_+d|_,dS)N)seedzRank z< convolutions are not currently implemented. Received: rank= kernel_sizestridesT allow_zero dilation_rate?rndim)-super__init__rrrandom SeedGeneratorseed_generatorrank ValueErrorfiltersrstandardize_tuplerrstandardize_paddingpaddingstandardize_data_format data_formatrrget activationrecurrent_activationuse_biasrkernel_initializerrecurrent_initializerbias_initializerunit_forget_biasrkernel_regularizerrecurrent_regularizerbias_regularizerrkernel_constraintrecurrent_constraintbias_constraintminmaxdropoutrecurrent_dropoutdropout_mask_countr input_spec state_size)selfr+r-rrr0r2rr4r5r6r7r8r9r:r;r<r=r>r?r@rCrDrkwargs __class__r%Y/var/www/html/chatgem/venv/lib/python3.10/site-packages/keras/src/layers/rnn/conv_lstm.pyr'UsP                zConvLSTMCell.__init__csFjdkrd}|dd_n d}|dd_djvr$td|||dur1td|tjdd|ddd _||}|_j|jd f_ jjjd f}j j j d j j d _j |jd jjd _jrjrfdd}nj}j jd fd|jjd_dSd_dS)Nchannels_firstrr!r$zvConvLSTM layers only support static input shapes for the spatial dimension. Received invalid input shape: input_shape=zvThe channel dimension of the inputs (last axis) should be defined. Found None. Full input shape received: input_shape=rN)r#shaper kernel)rO initializername regularizer constraintrecurrent_kernelcs`tjjfg|Ri|tdjfg|Ri|jjdfg|Ri|gS)Nonesr!)r concatenater9r-rr3)_argsrIrHr%rLr9s0 z,ConvLSTMCell.build..bias_initializerbias)rOrRrQrSrT)r2 spatial_dimsr,r r+rF input_dimrr- kernel_shape add_weightr7r;r>rPr8r<r?rUr6r:r9r=r@r[)rH inputs_shape states_shape channel_axisr]recurrent_kernel_shaper9r%rZrLbuildsj     zConvLSTMCell.buildFc)Cs,|d}|d}|r5d|jkrdkr5nn||}||d}||d}||d} ||d} n|}|}|} |} |rjd|jkrJdkrjnn||} || d} || d} || d}|| d}n|} |} |}|}tj|jd|jdd\}}}}tj|jd|jdd\}}}}|j rt|j d\}}}}nd \}}}}|j ||||j d }|j ||||j d }|j | |||j d }|j | |||j d }| | |} | | |}!| ||}"| ||}#||| }$|||!}%|%||$|||"}&|||#}'|'||&}(|(|(|&gfS) Nrrrrr!rr )axis)NNNN)r0)rCget_dropout_maskrDget_recurrent_dropout_maskrsplitrPr+rUr6r[ input_convr0recurrent_convr5r4))rHinputsstatestrainingh_tm1c_tm1dp_maskinputs_iinputs_finputs_cinputs_o rec_dp_maskh_tm1_ih_tm1_fh_tm1_ch_tm1_okernel_ikernel_fkernel_ckernel_orecurrent_kernel_irecurrent_kernel_frecurrent_kernel_crecurrent_kernel_obias_ibias_fbias_cbias_ox_ix_fx_cx_oh_ih_fh_ch_oifcohr%r%rLcalls`              zConvLSTMCell.callc Cs0tj||j|j|j|j|j|jd}|||gfS)Nrr0r2r)r compute_conv_output_shaper-rrr0r2r)rHr`raconv_output_shaper%r%rLcompute_output_shape$s z!ConvLSTMCell.compute_output_shapecCs^|jdkr|f|j|jf}n||jf|j}||d}tj||jdtj||jdgS)N channels_lastr)dtype)r2r\r]rrr compute_dtype)rH batch_size input_shape state_shaper%r%rLget_initial_state0s zConvLSTMCell.get_initial_statecCsrtj|||j||j|jd}|dur7|jdkr#d|jd|jf}n d|jfd|j}t||}||7}|S)Nrr)rr)rconvrr2rr+r-reshape)rHxwbr0conv_out bias_shaper[r%r%rLri;s  zConvLSTMCell.input_convcCs.tjd|jddd}tj|||d|jd}|S)NrrTrsame)rr0r2)rr.r+rrr2)rHrrrrr%r%rLrjMs  zConvLSTMCell.recurrent_convcsid|jd|jd|jd|jd|jd|jdt|jdt|j d |j d t |j d t |j d t |jd |jdt|jdt|jdt|jdt|jt|jt|j|j|j|jd}t}i||S)Nr-rrr0r2rr4r5r6r7r8r9r:r;r<r=r>)r?r@rCrDr)r-rrr0r2rr serializer4r5r6rr7r8r9r:rr;r<r=rr>r?r@rCrDrr& get_configrHconfig base_configrJr%rLrVsn        $ zConvLSTMCell.get_config)rrNrrrTrrrTNNNNNNrrNrN)F)Nr) __name__ __module__ __qualname____doc__r'rdrrrrirjr __classcell__r%r%rJrLrs<I G L <  rcseZdZdZ                 dCfd d ZdDfdd ZdEddZddZeddZ eddZ eddZ eddZ eddZ edd Zed!d"Zed#d$Zed%d&Zed'd(Zed)d*Zed+d,Zed-d.Zed/d0Zed1d2Zed3d4Zed5d6Zed7d8Zed9d:Zed;d<Zed=d>Zfd?d@ZedAdBZ Z!S)FConvLSTMaSAbstract 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 stdid|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|dd|_dS)Nr+r-rrr0r2rr4r5r6r7r8r9r:r;r<r=r>r?r@rCrDrrRconv_lstm_cellr)return_sequences return_state go_backwardsstatefulrr"r%)rr3r&r'r rF)rHr+r-rrr0r2rr4r5r6r7r8r9r:r;r<r=r>r?r@rCrDrrrrrrIcellrJr%rLr's|      zConvLSTM.__init__cstj||||dS)N) initial_statemaskrm)r&r)rH sequencesrrrmrJr%rLrsz ConvLSTM.callc Csv|d}|d}|f|dd}|j|ddd}|jr'||f|}n|f|}|jr9|f|}|||fS|S)Nrrr!)rrrr) rHsequences_shapeinitial_state_shapersteps step_shaper output_shapebatched_state_shaper%r%rLrs   zConvLSTM.compute_output_shapecCs8t|d}|jr |nd}|jrddg}|g|S|S)Nr)rflattenrr)rHrXr output_mask state_maskr%r%rL compute_mask,s  zConvLSTM.compute_maskcC|jjSrN)rr-rZr%r%rLr-5zConvLSTM.filterscCrrN)rrrZr%r%rLr9rzConvLSTM.kernel_sizecCrrN)rrrZr%r%rLr=rzConvLSTM.stridescCrrN)rr0rZr%r%rLr0ArzConvLSTM.paddingcCrrN)rr2rZr%r%rLr2ErzConvLSTM.data_formatcCrrN)rrrZr%r%rLrIrzConvLSTM.dilation_ratecCrrN)rr4rZr%r%rLr4MrzConvLSTM.activationcCrrN)rr5rZr%r%rLr5QrzConvLSTM.recurrent_activationcCrrN)rr6rZr%r%rLr6UrzConvLSTM.use_biascCrrN)rr7rZr%r%rLr7YrzConvLSTM.kernel_initializercCrrN)rr8rZr%r%rLr8]rzConvLSTM.recurrent_initializercCrrN)rr9rZr%r%rLr9arzConvLSTM.bias_initializercCrrN)rr:rZr%r%rLr:erzConvLSTM.unit_forget_biascCrrN)rr;rZr%r%rLr;irzConvLSTM.kernel_regularizercCrrN)rr<rZr%r%rLr<mrzConvLSTM.recurrent_regularizercCrrN)rr=rZr%r%rLr=qrzConvLSTM.bias_regularizercCrrN)rr>rZr%r%rLr>urzConvLSTM.kernel_constraintcCrrN)rr?rZr%r%rLr?yrzConvLSTM.recurrent_constraintcCrrN)rr@rZr%r%rLr@}rzConvLSTM.bias_constraintcCrrN)rrCrZr%r%rLrCrzConvLSTM.dropoutcCrrN)rrDrZr%r%rLrDrzConvLSTM.recurrent_dropoutcsid|jd|jd|jd|jd|jd|jdt|jdt|j d |j d t |j d t |j d t |jd |jdt|jdt|jdt|jdt|jt|jt|jt|j|j|j|jjd}t}|d=i||S)Nr-rrr0r2rr4r5r6r7r8r9r:r;r<r=activity_regularizer)r>r?r@rCrDrr)r-rrr0r2rrrr4r5r6rr7r8r9r:rr;r<r=rrr>r?r@rCrDrrr&rrrJr%rLrsv        ' zConvLSTM.get_configcCs|di|S)Nr%r%)clsrr%r%rL from_configszConvLSTM.from_config)rrNrrrTrrrTNNNNNNrrNFFFF)NNFrN)"rrrrr'rrrpropertyr-rrr0r2rr4r5r6r7r8r9r:r;r<r=r>r?r@rCrDr classmethodrrr%r%rJrLrsUD                       ,rN) keras.srcrrrrrrrkeras.src.layers.input_specr keras.src.layers.layerr %keras.src.layers.rnn.dropout_rnn_cellr keras.src.layers.rnn.rnnr keras.src.opsr keras.src.utilsrrrr%r%r%rLs"            q