2VhlDddlmZddlmZddlmZddlmZddlmZddlmZddlmZddl m Z dd l m Z dd l mZdd lmZed Gd de eZedGddeZy)) activations)backend) constraints) initializers)ops) regularizers) keras_export) InputSpec)Layer)DropoutRNNCell)RNNzkeras.layers.SimpleRNNCellc`eZdZdZ dfd ZdZddZd dZfdZxZ S) SimpleRNNCellav Cell class for SimpleRNN. This class processes one step within the whole time sequence input, whereas `keras.layer.SimpleRNN` processes the whole sequence. Args: units: Positive integer, dimensionality of the output space. activation: Activation function to use. Default: hyperbolic tangent (`tanh`). If you pass `None`, no activation is applied (ie. "linear" activation: `a(x) = x`). 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"`. 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`. 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: sequence: A 2D tensor, with shape `(batch, features)`. states: A 2D tensor with shape `(batch, units)`, which is the state from the previous time step. 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. Example: ```python inputs = np.random.random([32, 10, 8]).astype(np.float32) rnn = keras.layers.RNN(keras.layers.SimpleRNNCell(4)) output = rnn(inputs) # The output has shape `(32, 4)`. rnn = keras.layers.RNN( keras.layers.SimpleRNNCell(4), return_sequences=True, return_state=True ) # whole_sequence_output has shape `(32, 10, 4)`. # final_state has shape `(32, 4)`. whole_sequence_output, final_state = rnn(inputs) ``` c v|dkrtd|dt|di|||_tj j ||_||_tj||_ ||_ tj||_tj||_tj||_t#j||_t#j||_t#j| |_t+j| |_t+j| |_t+j| |_t3dt5d| |_t3dt5d||_|j|_|j|_y)NrzQReceived an invalid value for argument `units`, expected a positive integer, got .g?) ValueErrorsuper__init__seedrrandom SeedGeneratorseed_generatorunitsrget activationuse_biasrkernel_initializerrecurrent_initializerbias_initializerrkernel_regularizerrecurrent_regularizerbias_regularizerrkernel_constraintrecurrent_constraintbias_constraintminmaxdropoutrecurrent_dropout state_size output_size)selfrrrrr r!r"r#r$r%r&r'r*r+rkwargs __class__s O/home/dcms/DCMS/lib/python3.12/site-packages/keras/src/layers/rnn/simple_rnn.pyrzSimpleRNNCell.__init__Osh& A:4497!=  "6" %nn::4@ %//*5  "."2"23E"F%1%5%56K%L" , 0 01A B"."2"23E"F%1%5%56K%L" , 0 01A B!,1B!C$/OO4H$I!*?3C 12 !$S#c3D*E!F**::c|j|d|jfd|j|j|j|_|j|j|jfd|j |j|j|_ |jrE|j|jfd|j|j|j|_yd|_y)Nkernel)shapename initializer regularizer constraintrecurrent_kernelbias) add_weightrrr"r%r5r r#r&r;rr!r$r'r<)r. input_shapes r1buildzSimpleRNNCell.buildsoor?DJJ/////-- &  !%::tzz*#222200 !0!  ==zzm 11 11// (DIDIr2ct|ttfr|dn|}|j|}|j |}|r|||z}t j ||j}|j||jz }|r|||z}|t j ||jz}|j|j|}t|ttfr|gn|} || fS)Nr) isinstancelisttupleget_dropout_maskget_recurrent_dropout_maskrmatmulr5r<r;r) r.sequencestatestraining prev_outputdp_mask rec_dp_maskhoutput new_states r1callzSimpleRNNCell.calls#-ftUm#DfQi& ''155kB +')H JJx - 99 NA  /% 3KSZZ T-B-BCC ?? &__V,F *6D%= AVHv y  r2c^tj||jf|jgS)N)dtype)rzerosr, compute_dtype)r. batch_sizes r1get_initial_statezSimpleRNNCell.get_initial_states) IIz4??34;M;M N  r2c|jtj|j|jt j|j t j|jt j|jtj|jtj|jtj|jtj|jtj|jtj|j |j"|j$|j&d}t(|U}i||S)N)rrrrr r!r"r#r$r%r&r'r*r+r)rr serializerrrrr r!rr"r#r$rr%r&r'r*r+rr get_configr.config base_configr0s r1rYzSimpleRNNCell.get_configs-ZZ%//@ "."8"8''#&2%;%;**&!- 6 6t7L7L M"."8"8''#&2%;%;**&!- 6 6t7L7L M!,!6!6t7M7M!N$/$9$9))% +44T5I5IJ||!%!7!7II3 6g(* (+(((r2)tanhTglorot_uniform orthogonalrSNNNNNNrrN)FN) __name__ __module__ __qualname____doc__rr?rPrVrY __classcell__r0s@r1rrsT=D+* "! !/&b4!( ))r2rzkeras.layers.SimpleRNNcReZdZdZ dfd Zdfd ZedZedZedZ edZ edZ ed Z ed Z ed Zed Zed ZedZedZedZedZfdZedZxZS) SimpleRNNaFully-connected RNN where the output is to be fed back as the new input. Args: units: Positive integer, dimensionality of the output space. activation: Activation function to use. Default: hyperbolic tangent (`tanh`). If you pass None, no activation is applied (ie. "linear" activation: `a(x) = x`). use_bias: Boolean, (default `True`), whether the layer uses 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"`. 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. 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 and return the reversed sequence. stateful: Boolean (default: `False`). If `True`, the last state for each sample at index i in a batch will be used as the initial state for the sample of index i in the following batch. unroll: Boolean (default: `False`). If `True`, the network will be unrolled, else a symbolic loop will be used. Unrolling can speed-up an RNN, although it tends to be more memory-intensive. Unrolling is only suitable for short sequences. Call arguments: sequence: A 3D tensor, with shape `[batch, timesteps, feature]`. mask: Binary tensor of shape `[batch, timesteps]` indicating whether a given timestep should be masked. An individual `True` entry indicates that the corresponding timestep should be utilized, while a `False` entry indicates that the corresponding timestep should be ignored. training: Python boolean indicating whether the layer should behave in training mode or in inference mode. This argument is passed to the cell when calling it. This is only relevant if `dropout` or `recurrent_dropout` is used. initial_state: List of initial state tensors to be passed to the first call of the cell. Example: ```python inputs = np.random.random((32, 10, 8)) simple_rnn = keras.layers.SimpleRNN(4) output = simple_rnn(inputs) # The output has shape `(32, 4)`. simple_rnn = keras.layers.SimpleRNN( 4, return_sequences=True, return_state=True ) # whole_sequence_output has shape `(32, 10, 4)`. # final_state has shape `(32, 4)`. whole_sequence_output, final_state = simple_rnn(inputs) ``` c t|fid|d|d|d|d|d|d|d| d | d | d | d |d |d|d|jddd|jdddd}t| |f|||||d|t dg|_y)Nrrrr r!r"r#r$r%r&r'r*r+rrR trainableTr7simple_rnn_cell)return_sequences return_state go_backwardsstatefulunroll)ndim)rrrrr input_spec)r.rrrrr r!r"r#r$activity_regularizerr%r&r'r*r+rlrmrnrorprr/cellr0s r1rzSimpleRNN.__init__(s2  !   2  #8  .   2 #8 . 0 "6 ,  0  **Wd+! "jjd3# $#% (   -%%   %!,-r2c*t|||||S)N)maskrI initial_state)rrP)r. sequencesrxrwrIr0s r1rPzSimpleRNN.call`s"w| D8=  r2c.|jjSr`)rurr.s r1rzSimpleRNN.unitsesyyr2c.|jjSr`)rurr{s r1rzSimpleRNN.activationisyy###r2c.|jjSr`)rurr{s r1rzSimpleRNN.use_biasmsyy!!!r2c.|jjSr`)rurr{s r1rzSimpleRNN.kernel_initializerqyy+++r2c.|jjSr`)rur r{s r1r zSimpleRNN.recurrent_initializeruyy...r2c.|jjSr`)rur!r{s r1r!zSimpleRNN.bias_initializeryyy)))r2c.|jjSr`)rur"r{s r1r"zSimpleRNN.kernel_regularizer}rr2c.|jjSr`)rur#r{s r1r#zSimpleRNN.recurrent_regularizerrr2c.|jjSr`)rur$r{s r1r$zSimpleRNN.bias_regularizerrr2c.|jjSr`)rur%r{s r1r%zSimpleRNN.kernel_constraintyy***r2c.|jjSr`)rur&r{s r1r&zSimpleRNN.recurrent_constraintsyy---r2c.|jjSr`)rur'r{s r1r'zSimpleRNN.bias_constraintsyy(((r2c.|jjSr`)rur*r{s r1r*zSimpleRNN.dropoutsyy   r2c.|jjSr`)rur+r{s r1r+zSimpleRNN.recurrent_dropoutrr2c$|jtj|j|jt j|j t j|jt j|jtj|jtj|jtj|jtj|jtj|jtj|j tj|j"|j$|j&d}t(|U}|d=i||S)N)rrrrr r!r"r#r$rtr%r&r'r*r+ru)rrrXrrrrr r!rr"r#r$rtrr%r&r'r*r+rrYrZs r1rYzSimpleRNN.get_configsGZZ%//@ "."8"8''#&2%;%;**&!- 6 6t7L7L M"."8"8''#&2%;%;**&!- 6 6t7L7L M$0$:$:))%"-!6!6t7M7M!N$/$9$9))% +44T5I5IJ||!%!7!77 :g(*  (+(((r2c|di|S)Nrr)clsr[s r1 from_configzSimpleRNN.from_configs}V}r2)r]Tr^r_rSNNNNNNNrrFFFFFN)NNF)rarbrcrdrrPpropertyrrrrr r!r"r#r$r%r&r'r*r+rY classmethodrrerfs@r1rhrhsQl+* "!! -6.p $$"",,//**,,//**++..))!!++ )Dr2rhN) keras.srcrrrrrrkeras.src.api_exportr keras.src.layers.input_specr keras.src.layers.layerr %keras.src.layers.rnn.dropout_rnn_cellr keras.src.layers.rnn.rnnr rrhrr2r1rsq!!""-1(@(*+A)E>A),A)H&'mm(mr2