&Vf dZddlZddlmZddlmZddlmZddlmZddlm Z ddl m Z dd l m Z dd lmZdd lmZGd d eZeddgdZGddeZeddgdZGddeZeddgdZGddeZeddgdZGd d!eZed"d#gd$ZGd%d&eZegd'd(ZGd)d*eZed+d,gd-ZGd.d/eZ ed0d1gdd3Z!Gd4d5eZ"ed6d7gd8Z#Gd9d:eZ$egd;d<Z%Gd=d>eZ&ed?d@gddBZ'GdCdDeZ(edEdFgdGZ)GdHdIeZ*edJdKgddMZ+GdNdOeZ,edPdQgddSZ-GdTdUeZ.edVdWgddXZ/GdYdZeZ0ed[d\g dd^Z1Gd_d`eZ2edadbg ddcZ3GdddeeZ4edfdgg ddiZ5GdjdkeZ6edldmg ddnZ7GdodpeZ8edqdrg ddsZ9GdtdueZ:edvdwg ddxZ;GdydzeZ<ed{d|gdd~Z=GddeZ>eddgddZ?GddeZ@eddgddZAGddeZBeddgddZCGddeZDeddg ddZEGddeZFeddgddZGGddeZHeddg ddZIGddeZJeddgddZKGddeZLeddg ddZMGddeZNeddgddZOddZPGddeZQeddgdZRdS)z>Commonly-used neural network operations not included in NumPy.N)backend) keras_export) KerasTensor)any_symbolic_tensors)standardize_data_format)#compute_conv_transpose_output_shape)operation_utils) Operation) reduce_shapeceZdZdZdZdS)Reluc@tj|SN)rnnreluselfxs M/var/www/html/software/conda/lib/python3.11/site-packages/keras/src/ops/nn.pycallz Relu.callzq!!!c8t|j|jSNdtypershaperrs rcompute_output_speczRelu.compute_output_spec17!'2222rN__name__ __module__ __qualname__rrrrr r 2"""33333rr zkeras.ops.reluzkeras.ops.nn.reluct|fr!t|Stj|S)aQRectified linear unit activation function. It is defined as `f(x) = max(0, x)`. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x1 = keras.ops.convert_to_tensor([-1.0, 0.0, 1.0, 0.2]) >>> keras.ops.relu(x1) array([0.0, 0.0, 1.0, 0.2], dtype=float32) )rr symbolic_callrrrrs rrrsB$QD!!'vv##A&&& :??1  rceZdZdZdZdS)Relu6c@tj|Sr)rrrelu6rs rrz Relu6.call2sz"""rc8t|j|jSrrrs rrzRelu6.compute_output_spec5r rNr!r%rrr+r+1s2###33333rr+zkeras.ops.relu6zkeras.ops.nn.relu6ct|fr!t|Stj|S)aRectified linear unit activation function with upper bound of 6. It is defined as `f(x) = np.clip(x, 0, 6)`. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = keras.ops.convert_to_tensor([-3.0, -2.0, 0.1, 0.2, 6.0, 8.0]) >>> keras.ops.relu6(x) array([0.0, 0.0, 0.1, 0.2, 6.0, 6.0], dtype=float32) )rr+r(rrr-r)s rr-r-9sD$QD!!(ww$$Q''' :  A  rceZdZdZdZdS)Sigmoidc@tj|Sr)rrsigmoidrs rrz Sigmoid.callQsz!!!$$$rc8t|j|jSrrrs rrzSigmoid.compute_output_specTr rNr!r%rrr1r1Ps2%%%33333rr1zkeras.ops.sigmoidzkeras.ops.nn.sigmoidct|fr!t|Stj|S)apSigmoid activation function. It is defined as `f(x) = 1 / (1 + exp(-x))`. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = keras.ops.convert_to_tensor([-6.0, 1.0, 0.0, 1.0, 6.0]) >>> keras.ops.sigmoid(x) array([0.00247262, 0.7310586, 0.5, 0.7310586, 0.9975274], dtype=float32) )rr1r(rrr3r)s rr3r3XsD&QD!!*yy&&q))) :  a  rceZdZdZdZdS)Softplusc@tj|Sr)rrsoftplusrs rrz Softplus.callqz""1%%%rc8t|j|jSrrrs rrzSoftplus.compute_output_spectr rNr!r%rrr7r7p2&&&33333rr7zkeras.ops.softpluszkeras.ops.nn.softplusct|fr!t|Stj|S)aSoftplus activation function. It is defined as `f(x) = log(exp(x) + 1)`, where `log` is the natural logarithm and `exp` is the exponential function. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = keras.ops.convert_to_tensor([-0.555, 0.0, 0.555]) >>> keras.ops.softplus(x) array([0.45366603, 0.6931472, 1.008666], dtype=float32) )rr7r(rrr9r)s rr9r9xsD(QD!!+zz''*** :  q ! !!rceZdZdZdZdS)Softsignc@tj|Sr)rrsoftsignrs rrz Softsign.callr:rc8t|j|jSrrrs rrzSoftsign.compute_output_specr rNr!r%rrr?r?r<rr?zkeras.ops.softsignzkeras.ops.nn.softsignct|fr!t|Stj|S)asSoftsign activation function. It is defined as `f(x) = x / (abs(x) + 1)`. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = keras.ops.convert_to_tensor([-0.100, -10.0, 1.0, 0.0, 100.0]) >>> keras.ops.softsign(x) Array([-0.09090909, -0.90909094, 0.5, 0.0, 0.990099], dtype=float32) )rr?r(rrrAr)s rrArAsD&QD!!+zz''*** :  q ! !!rceZdZdZdZdS)Siluc@tj|Sr)rrsilurs rrz Silu.callrrc8t|j|jSrrrs rrzSilu.compute_output_specr rNr!r%rrrErEr&rrE)zkeras.ops.siluzkeras.ops.nn.siluzkeras.ops.swishzkeras.ops.nn.swishct|fr!t|Stj|S)aZSigmoid Linear Unit (SiLU) activation function, also known as Swish. The SiLU activation function is computed by the sigmoid function multiplied by its input. It is defined as `f(x) = x * sigmoid(x)`. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = keras.ops.convert_to_tensor([-6.0, 1.0, 0.0, 1.0, 6.0]) >>> keras.ops.sigmoid(x) array([0.00247262, 0.7310586, 0.5, 0.7310586, 0.9975274], dtype=float32) >>> keras.ops.silu(x) array([-0.0148357, 0.7310586, 0.0, 0.7310586, 5.9851646], dtype=float32) )rrEr(rrrGr)s rrGrGsB:QD!!'vv##A&&& :??1  rceZdZdZdZdS) LogSigmoidc@tj|Sr)rr log_sigmoidrs rrzLogSigmoid.callsz%%a(((rc8t|j|jSrrrs rrzLogSigmoid.compute_output_specr rNr!r%rrrKrKs2)))33333rrKzkeras.ops.log_sigmoidzkeras.ops.nn.log_sigmoidct|fr!t|Stj|S)aLogarithm of the sigmoid activation function. It is defined as `f(x) = log(1 / (1 + exp(-x)))`. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = keras.ops.convert_to_tensor([-0.541391, 0.0, 0.50, 5.0]) >>> keras.ops.log_sigmoid(x) array([-1.0000418, -0.6931472, -0.474077, -0.00671535], dtype=float32) )rrKr(rrrMr)s rrMrMsD0QD!!-||))!,,, : ! !! $ $$rc,eZdZdfd ZdZdZxZS) LeakyRelu皙?cVt||_dSr)super__init__negative_slope)rrV __class__s rrUzLeakyRelu.__init__s' ,rcLtj||jSr)rr leaky_relurVrs rrzLeakyRelu.callsz$$Q(;<<= 0`. Args: x: Input tensor. negative_slope: Slope of the activation function at x < 0. Defaults to `0.2`. Returns: A tensor with the same shape as `x`. Example: >>> x = np.array([-1., 0., 1.]) >>> x_leaky_relu = keras.ops.leaky_relu(x) >>> print(x_leaky_relu) array([-0.2, 0. , 1. ], shape=(3,), dtype=float64) )rV)rrQr(rrrY)rrVs rrYrY sM0QD!!:((66q999 : > B BBrceZdZdZdZdS) HardSigmoidc@tj|Sr)rr hard_sigmoidrs rrzHardSigmoid.call*sz&&q)))rc8t|j|jSrrrs rrzHardSigmoid.compute_output_spec-r rNr!r%rrrara)s2***33333rrazkeras.ops.hard_sigmoidzkeras.ops.nn.hard_sigmoidct|fr!t|Stj|S)aHard sigmoid activation function. It is defined as: `0 if x < -2.5`, `1 if x > 2.5`, `(0.2 * x) + 0.5 if -2.5 <= x <= 2.5`. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = np.array([-1., 0., 1.]) >>> x_hard_sigmoid = keras.ops.hard_sigmoid(x) >>> print(x_hard_sigmoid) array([0.3, 0.5, 0.7], shape=(3,), dtype=float64) )rrar(rrrcr)s rrcrc1sD6QD!!.}}**1--- : " "1 % %%rceZdZdZdZdS)HardSiluc@tj|Sr)rr hard_silurs rrz HardSilu.callRsz##A&&&rc8t|j|jSrrrs rrzHardSilu.compute_output_specUr rNr!r%rrrgrgQs2'''33333rrg)zkeras.ops.hard_siluzkeras.ops.nn.hard_siluzkeras.ops.hard_swishzkeras.ops.nn.hard_swishct|fr!t|Stj|S)aHard SiLU activation function, also known as Hard Swish. It is defined as: - `0` if `if x < -3` - `x` if `x > 3` - `x * (x + 3) / 6` if `-3 <= x <= 3` It's a faster, piecewise linear approximation of the silu activation. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = keras.ops.convert_to_tensor([-3.0, -1.0, 0.0, 1.0, 3.0]) >>> keras.ops.hard_silu(x) array([-0.0, -0.3333333, 0.0, 0.6666667, 3.0], shape=(5,), dtype=float32) )rrgr(rrrir)s rririYsE@QD!!+zz''*** :   " ""rc,eZdZdfd ZdZdZxZS)Elu?cVt||_dSr)rTrUalpha)rrprWs rrUz Elu.__init__s$  rcNtj||jS)Nrp)rrelurprs rrzElu.callsz~~atz~222rc8t|j|jSrrrs rrzElu.compute_output_specr rrnr\r^s@rrmrm~s[3333333333rrmz keras.ops.eluzkeras.ops.nn.elurnct|fr"t||Stj||S)aExponential Linear Unit activation function. It is defined as: `f(x) = alpha * (exp(x) - 1.) for x < 0`, `f(x) = x for x >= 0`. Args: x: Input tensor. alpha: A scalar, slope of positive section. Defaults to `1.0`. Returns: A tensor with the same shape as `x`. Example: >>> x = np.array([-1., 0., 1.]) >>> x_elu = keras.ops.elu(x) >>> print(x_elu) array([-0.63212055, 0., 1.], shape=(3,), dtype=float64) rr)rrmr(rrrs)rrps rrsrssH.QD!!+5zz''*** :>>!5> ) ))rceZdZdZdZdS)Seluc@tj|Sr)rrselurs rrz Selu.callrrc8t|j|jSrrrs rrzSelu.compute_output_specr rNr!r%rrrxrxr&rrxzkeras.ops.seluzkeras.ops.nn.seluct|fr!t|Stj|S)aScaled Exponential Linear Unit (SELU) activation function. It is defined as: `f(x) = scale * alpha * (exp(x) - 1.) for x < 0`, `f(x) = scale * x for x >= 0`. Args: x: Input tensor. Returns: A tensor with the same shape as `x`. Example: >>> x = np.array([-1., 0., 1.]) >>> x_selu = keras.ops.selu(x) >>> print(x_selu) array([-1.11133055, 0., 1.05070098], shape=(3,), dtype=float64) )rrxr(rrrzr)s rrzrzsB.QD!!'vv##A&&& :??1  rc,eZdZdfd ZdZdZxZS)GeluTcVt||_dSr)rTrU approximate)rrrWs rrUz Gelu.__init__' &rcLtj||jSr)rrgelurrs rrz Gelu.callszq$"2333rc8t|j|jSrrrs rrzGelu.compute_output_specr rTr\r^s@rr~r~s[''''''4443333333rr~zkeras.ops.geluzkeras.ops.nn.geluTct|fr"t||Stj||S)aGaussian Error Linear Unit (GELU) activation function. If `approximate` is `True`, it is defined as: `f(x) = 0.5 * x * (1 + tanh(sqrt(2 / pi) * (x + 0.044715 * x^3)))` Or if `approximate` is `False`, it is defined as: `f(x) = x * P(X <= x) = 0.5 * x * (1 + erf(x / sqrt(2)))`, where `P(X) ~ N(0, 1)`. Args: x: Input tensor. approximate: Approximate version of GELU activation. Defaults to `True`. Returns: A tensor with the same shape as `x`. Example: >>> x = np.array([-1., 0., 1.]) >>> x_gelu = keras.ops.gelu(x) >>> print(x_gelu) array([-0.15865525, 0., 0.84134475], shape=(3,), dtype=float64) )rr~r(rrr)rrs rrrsH4QD!!2K  ..q111 :??1k * **rc,eZdZdfd ZdZdZxZS)SoftmaxcVt||_dSrrTrUaxisrrrWs rrUzSoftmax.__init__$  rcNtj||jSNr)rrsoftmaxrrs rrz Softmax.callsz!!!$)!444rc8t|j|jSrrrs rrzSoftmax.compute_output_specr rrr\r^s@rrrs[5553333333rrzkeras.ops.softmaxzkeras.ops.nn.softmaxrcLt|tr1|j|dkr tjd|d|jdt |fr"t ||St|tr|j}g}t|}d}|t|krg||vr0d}||vr|||z}|dz }||v| |n | |||dz }|t|kgtj ||}tj|d}tj ||}|Stj||S)aSoftmax activation function. The elements of the output vector lie within the range `(0, 1)`, and their total sum is exactly 1 (excluding the floating point rounding error). Each vector is processed independently. The `axis` argument specifies the axis along which the function is applied within the input. It is defined as: `f(x) = exp(x) / sum(exp(x))` Args: x: Input tensor. axis: Integer, axis along which the softmax is applied. Returns: A tensor with the same shape as `x`. Example: >>> x = np.array([-1., 0., 1.]) >>> x_softmax = keras.ops.softmax(x) >>> print(x_softmax) array([0.09003057, 0.24472847, 0.66524096], shape=(3,), dtype=float64) z"You are using a softmax over axis z of a tensor of shape z. This axis has size 1. The softmax operation will always return the value 1, which is likely not what you intended. Did you mean to use a sigmoid instead?rrr) isinstanceintrwarningswarnrrr(tuplesetlenappendrnumpyreshaperrrroriginal_shape new_shape skip_dimsisizes rrrs>$ !!3!3  5 5 5$%G 5 5 5   QD!!.t}}**1---$0 II #n%%%%I~~9nnN1--DFA9nn  &&&&  !2333Q#n%%%% M ! !!Y / / J  qr  * * M ! !!^ 4 4z!!!$!///rc,eZdZdfd ZdZdZxZS) LogSoftmaxrcVt||_dSrrrs rrUzLogSoftmax.__init__BrrcNtj||jSr)rr log_softmaxrrs rrzLogSoftmax.callFsz%%adi%888rc8t|j|jSrrrs rrzLogSoftmax.compute_output_specIr rrr\r^s@rrrAs[9993333333rrzkeras.ops.log_softmaxzkeras.ops.nn.log_softmaxct|fr"t||St|tr|j}g}t |}d}|t|krg||vr0d}||vr|||z}|dz }||v||n ||||dz }|t|kgtj ||}tj |d}tj ||}|Stj ||S)aLog-softmax activation function. It is defined as: `f(x) = x - max(x) - log(sum(exp(x - max(x))))` Args: x: Input tensor. axis: Integer, axis along which the log-softmax is applied. Defaults to `-1`. Returns: A tensor with the same shape as `x`. Example: >>> x = np.array([-1., 0., 1.]) >>> x_log_softmax = keras.ops.log_softmax(x) >>> print(x_log_softmax) array([-2.40760596, -1.40760596, -0.40760596], shape=(3,), dtype=float64) rrrr)rrr(rrrrrrrrrrrrs rrrMsg8QD!!1$--a000$4 II #n%%%%I~~9nnN1--DFA9nn  &&&&  !2333Q#n%%%% M ! !!Y / / J " "12 " . . M ! !!^ 4 4z%%ad%333rc2eZdZ dfd ZdZdZxZS)MaxPoolNvalidct||_||_||_||_dSrrTrU pool_sizestrideslowerpadding data_formatrrrrrrWs rrUzMaxPool.__init__F " }} &rcptj||j|j|j|jSr)rrmax_poolrrrrrinputss rrz MaxPool.calls4z""  N L L      rctj|j|j|j|j|j}t||jSr r compute_pooling_output_shaperrrrrrrrr output_shapes rrzMaxPool.compute_output_specE&C L N L L     >> x = keras.ops.convert_to_tensor([1, 3, 2, 0]) >>> one_hot(x, num_classes=4) array([[0. 1. 0. 0.] [0. 0. 0. 1.] [0. 0. 1. 0.] [1. 0. 0. 0.]], shape=(4, 4), dtype=float32) r)rrr(rrr r)rrrrrs rr r sFQD!! d%   -    :    'w~''   rc,eZdZdfd ZdZdZxZS)BinaryCrossentropyFcVt||_dSr)rTrU from_logits)rrrWs rrUzBinaryCrossentropy.__init__9rrcPtj|||jS)Nr)rrbinary_crossentropyrrtargetoutputs rrzBinaryCrossentropy.call=s+z-- F(8.   rc|j|jkrtd|jd|jt|j|jS)NQArguments `target` and `output` must have the same shape. Received: target.shape=, output.shape=r)rrrrrs rrz&BinaryCrossentropy.compute_output_specBsd <6< ' 'L & LL=C\LL  6>> target = keras.ops.convert_to_tensor([0, 1, 1, 0]) >>> output = keras.ops.convert_to_tensor([0.1, 0.9, 0.8, 0.2]) >>> binary_crossentropy(target, output) array([0.10536054 0.10536054 0.22314355 0.22314355], shape=(4,), dtype=float32) r)rrr(rrr)rrrs rrrLsjLVV,-- !k:::HH F    : ) )K *  rc,eZdZdfd ZdZdZxZS)CategoricalCrossentropyFrcdt||_||_dSrrTrUrrrrrrWs rrUz CategoricalCrossentropy.__init__|, & rc\tj|||j|jSNrr)rrcategorical_crossentropyrrrs rrzCategoricalCrossentropy.calls/z22 F(8ty3   rc|j|jkrtd|jd|jt|jdkrtd|jd|jt|jdd|jS)Nr!r"rzPArguments `target` and `output` must be at least rank 1. Received: target.shape=rr)rrrrrrs rrz+CategoricalCrossentropy.compute_output_specs <6< ' 'L & LL=C\LL  v|  q L & LL=C\LL  6<,FLAAAArFrr\r^s@rr&r&{sb    B B B B B B Brr&z"keras.ops.categorical_crossentropyz%keras.ops.nn.categorical_crossentropyct||fr%t||||Stj||||S)aComputes categorical cross-entropy loss between target and output tensor. The categorical cross-entropy loss is commonly used in multi-class classification tasks where each input sample can belong to one of multiple classes. It measures the dissimilarity between the target and output probabilities or logits. Args: target: The target tensor representing the true categorical labels. Its shape should match the shape of the `output` tensor except for the last dimension. output: The output tensor representing the predicted probabilities or logits. Its shape should match the shape of the `target` tensor except for the last dimension. from_logits: (optional) Whether `output` is a tensor of logits or probabilities. Set it to `True` if `output` represents logits; otherwise, set it to `False` if `output` represents probabilities. Defaults to`False`. axis: (optional) The axis along which the categorical cross-entropy is computed. Defaults to `-1`, which corresponds to the last dimension of the tensors. Returns: Integer tensor: The computed categorical cross-entropy loss between `target` and `output`. Example: >>> target = keras.ops.convert_to_tensor( ... [[1, 0, 0], ... [0, 1, 0], ... [0, 0, 1]]) >>> output = keras.ops.convert_to_tensor( ... [[0.9, 0.05, 0.05], ... [0.1, 0.8, 0.1], ... [0.2, 0.3, 0.5]]) >>> categorical_crossentropy(target, output) array([0.10536054 0.22314355 0.6931472 ], shape=(3,), dtype=float32) r-)rr&r(rrr.rrrrs rr.r.so`VV,--(&#$   - ' ' ( : . .Kd /  rc,eZdZdfd ZdZdZxZS)SparseCategoricalCrossentropyFrcdt||_||_dSrr(r)s rrUz&SparseCategoricalCrossentropy.__init__r*rc\tj|||j|jSr,)rrsparse_categorical_crossentropyrrrs rrz"SparseCategoricalCrossentropy.calls/z99 F(8ty:   rct|jdkrtd|j|j}t|t|jkr|ddkr |dd}||jddkrtd|jd|jt|jdd|jS)NrzBArgument `output` must be at least rank 1. Received: output.shape=rzcArguments `target` and `output` must have the same shape up until the last dimension: target.shape=r"r)rrrrr)rrr target_shapes rrz1SparseCategoricalCrossentropy.compute_output_specs v|  q / & //  | |  FL 1 1 1 1l26F!6K6K',L 6<, , ,L & LL=C\LL  6<,FLAAAArr0r\r^s@rr4r4sb    BBBBBBBrr4z)keras.ops.sparse_categorical_crossentropyz,keras.ops.nn.sparse_categorical_crossentropyct||fr%t||||Stj||||S)aComputes sparse categorical cross-entropy loss. The sparse categorical cross-entropy loss is similar to categorical cross-entropy, but it is used when the target tensor contains integer class labels instead of one-hot encoded vectors. It measures the dissimilarity between the target and output probabilities or logits. Args: target: The target tensor representing the true class labels as integers. Its shape should match the shape of the `output` tensor except for the last dimension. output: The output tensor representing the predicted probabilities or logits. Its shape should match the shape of the `target` tensor except for the last dimension. from_logits: (optional) Whether `output` is a tensor of logits or probabilities. Set it to `True` if `output` represents logits; otherwise, set it to `False` if `output` represents probabilities. Defaults to`False`. axis: (optional) The axis along which the sparse categorical cross-entropy is computed. Defaults to `-1`, which corresponds to the last dimension of the tensors. Returns: Integer tensor: The computed sparse categorical cross-entropy loss between `target` and `output`. Example: >>> target = keras.ops.convert_to_tensor([0, 1, 2], dtype=int32) >>> output = keras.ops.convert_to_tensor( ... [[0.9, 0.05, 0.05], ... [0.1, 0.8, 0.1], ... [0.2, 0.3, 0.5]]) >>> sparse_categorical_crossentropy(target, output) array([0.10536056 0.22314355 0.6931472 ], shape=(3,), dtype=float32) r-)rr4r(rrr7r2s rr7r7so\VV,--(,#$   - ' ' ( : 5 5Kd 6  rc.eZdZ dfd ZdZdZxZS)MultiHotNrFc |d|vr|d}|tdtjdi|||_||_|pt j|_||_ dS)N num_tokens)Argument `num_classes` must be specified.r%) poprrTrUrrrrrr)rrrrrkwargsrWs rrUzMultiHot.__init__%s  <6#9#9 **\22K  HII I""6"""& .gn..  rcftj||j|j|jS)N)rrr)rr multi_hotrrrrs rrz MultiHot.call2s4z## (* $   rctt|dg}|jdkr||jnq|jdkr9|jt |kr!||j|jn-tdt |jd|jdt |dkr |dg}n|dg|ddz}t||j |j S) Nrrrr r r rr)rrrs rrzMultiHot.compute_output_spec:s wvw3344 9?? NN4+ , , , , Y!^^ CLL 8 8 NN49d&6 7 7 7 7)#fl2C2C)) I)))  w<<1  r{mGGqzlWQRR[0G7&,t{KKKKrNrNFr\r^s@rr<r<$se>> data = keras.ops.convert_to_tensor([0, 4]) >>> keras.ops.multi_hot(data, num_classes=5) array([1.0, 0.0, 0.0, 0.0, 1.0], dtype=float32) Nr>r?)r@rrr<r(rrrC)rrrrrrAs rrCrCNsD|v55jj.. DEEEVI&&P T5&99GGOOO :   T5& I IIrc,eZdZdfd ZdZdZxZS)MomentsFcrt||_||_||_dSr)rTrUaxeskeepdims synchronized)rrKrLrMrWs rrUzMoments.__init__|s5    (rcftj||j|j|jS)N)rKrLrM)rrmomentsrKrLrMrs rrz Moments.calls4z!! ]* "   rctt|j|j|j|jtt|j|j|j|jfS)N)rrLr)rr rrKrLrrs rrzMoments.compute_output_specsl QW49t}MMMg    QW49t}MMMg     rFFr\r^s@rrIrI{s[))))))            rrIzkeras.ops.momentszkeras.ops.nn.momentsct|fr%t||||Stj||||S)a!Calculates the mean and variance of `x`. The mean and variance are calculated by aggregating the contents of `x` across `axes`. If `x` is 1-D and `axes = [0]` this is just the mean and variance of a vector. Args: x: Input tensor. axes: A list of axes which to compute mean and variance. keepdims: If this is set to `True`, the axes which are reduced are left in the result as dimensions with size one. synchronized: Only applicable with the TensorFlow backend. If `True`, synchronizes the global batch statistics (mean and variance) across all devices at each training step in a distributed training strategy. If `False`, each replica uses its own local batch statistics. Returns: A tuple containing two tensors - mean and variance. Example: >>> x = keras.ops.convert_to_tensor([0, 1, 2, 3, 100], dtype="float32") >>> keras.ops.moments(x, axes=[0]) (array(21.2, dtype=float32), array(1553.3601, dtype=float32)) )rM)rrIr(rrrO)rrKrLrMs rrOrOsbDQD!! tXLAAAOO     :  axl  K KKrc*eZdZfdZdZdZxZS) BatchNormcdt||_||_dSr)rTrUrepsilon)rrrVrWs rrUzBatchNorm.__init__s+   rc D||krtd|d|d|ddS)N Arguments `z9` must be a vector of length `x.shape[axis]`. Expected: `z`. Received: `r r)rnamerexpected_shapes r _check_shapezBatchNorm._check_shapesW N " "'d''/=''#'''  # "rc|j|jf}|dt|j||dt|j||)|dt|j|||ur)|dt|j|t |j|jS)Nmeanvarianceoffsetscaler)rrr\rrr)rrr^r_r`rars rrzBatchNorm.compute_output_specs#% &% "3"3U;;; *eHN&;&;UCCC     hfl(;(;U C C C      guU['9'95 A A A17!'2222r)r"r#r$rUr\rr]r^s@rrTrTsV 3333333rrTzkeras.ops.batch_normalizationz keras.ops.nn.batch_normalizationMbP?c t|||||fr't|||||||Stj|||||||S)aNormalizes `x` by `mean` and `variance`. This op is typically used by the batch normalization step in a neural network. It normalizes the input tensor along the given axis. Args: x: Input tensor. mean: A mean vector of the same length as the `axis` dimension of the input thensor. variance: A variance vector of the same length as the `axis` dimension of the input tensor. axis: Integer, the axis that should be normalized. offset: An offset vector of the same length as the `axis` dimension of the input tensor. If not `None`, `offset` is added to the normalized tensor. Defaults to `None`. scale: A scale vector of the same length as the `axis` dimension of the input tensor. If not `None`, the normalized tensor is multiplied by `scale`. Defaults to `None`. epsilon: Small float added to variance to avoid dividing by zero. Defaults to 1e-3. Returns: The normalized tensor. Example: >>> x = keras.ops.convert_to_tensor( ... [[0.1, 0.2, 0.3], [0.4, 0.5, 0.6], [0.7, 0.8, 0.9]] ... ) >>> keras.ops.batch_normalization( ... x, ... mean=[0.4, 0.5, 0.6], ... variance=[0.67, 0.67, 0.67], ... axis=-1 ... ) array([[-3.6624e-01, -3.6624e-01, -3.6624e-01], [-4.6445e-09, 0.0000e+00, -1.8578e-08], [ 3.6624e-01, 3.6624e-01, 3.6624e-01]]) )rrTr(rrbatch_normalization)rr^r_rr`rarVs rrdrdsxbQh>?? w''55 tXvu    : ) ) 44  rc2eZdZdfd ZdZdZdZxZS)CTCLossrcVt||_dSr)rTrU mask_index)rrhrWs rrUzCTCLoss.__init__s$ $rcRtj|||||jSr)rrctc_lossrh)rrr target_length output_lengths rrz CTCLoss.calls)z"" FM=$/   rc b|d|dkrtd|d|d|d|d dS)NrrXz` and `z8` must have the same first dimension. Received shapes: `z`.rY)rname1shape1name2shape2s r_check_shape_first_dimzCTCLoss._check_shape_first_dimsk !9q ! !?e??E??%+??4:???  " !rc<|d|jd|j|d|jd|j|d|jd|jtj|jd}t |jdf|S)Nrrrkrlfloat32rr)rrrr result_typerr)rrrrkrlrs rrzCTCLoss.compute_output_spec's ## flHfl    ## ]0(FL    ## ]0(FL   #FL)<<FLO-U;;;;rr)r"r#r$rUrrrrr]r^s@rrfrfsj%%%%%%    < < < < < < >>>rrrE)r"r#r$rUrrr]r^s@rrrs[ ***???????rrzkeras.ops.normalizezkeras.ops.nn.normalizerEct|fr$t|||St|||S)aNormalizes `x` over the specified axis. It is defined as: `normalize(x) = x / max(norm(x), epsilon)`. Args: x: Input tensor. axis: The axis or axes along which to perform normalization. Default to -1. order: The exponent value in the norm formulation. Defaults to 2. Returns: The normalized array. Example: >>> x = keras.ops.convert_to_tensor([[1, 2, 3], [4, 5, 6]]) >>> x_norm = keras.ops.math.normalize(x) >>> print(x_norm) array([[0.26726124 0.5345225 0.8017837 ] [0.45584232 0.5698029 0.68376344]], shape=(2, 3), dtype=float32) r)rrr(r)rrrs r normalizersM<QD!!Bd%000>>qAAA ad% 0 0 00rct|tr|dkstd|tj|}t |jdkr!tj|d}tj }tj |||d}tj ||}tj ||S)Nrz6Argument `order` must be an int >= 1. Received: order=rrT)ordrrL)rrrrconvert_to_tensorrrr expand_dimsrVlinalgnormmaximumdivide)rrrrVrdenoms rrrs eS ! ! ! LU L L    !!$$A 17||q M % %aa % 0 0oG >  qe$  F FD M ! !$ 0 0E =  5 ) ))rc*eZdZfdZdZdZxZS)PSNRcVt||_dSr)rTrUmax_val)rrrWs rrUz PSNR.__init__s&  rcPtj|||jS)Nx1x2r)rrpsnrrrrrs rrz PSNR.calls*zL   rct|jt|jkrtdtdS)NzInputs must have the same rankr%r)rrrrrs rrzPSNR.compute_output_specs= rx==CMM ) )=>> >$$$$rr\r^s@rrrsV   %%%%%%%rrzkeras.ops.psnrzkeras.ops.nn.psnrct||fr#t|||Stj|||S)a:Peak Signal-to-Noise Ratio (PSNR) function. This function computes the Peak Signal-to-Noise Ratio between two signals, `x1` and `x2`. PSNR is a measure of the quality of a reconstructed signal. The higher the PSNR, the closer the reconstructed signal is to the original signal. Note that it can become negative when the signal power is smaller that the noise power. Args: x1: The first input signal. x2: The second input signal. Must have the same shape as `x1`. max_val: The maximum possible value in the signals. Returns: float: The PSNR value between `x1` and `x2`. Examples: >>> x1 = keras.random.normal((2, 4, 4, 3)) >>> x2 = keras.random.normal((2, 4, 4, 3)) >>> max_val = 1.0 >>> keras.ops.nn.psnr(x1, x2, max_val) -3.1697404 )rrr(rrrrs rrrskF       -B    :??    rr[rurrrrrrr#r0rFrQ)NNrbrvrr)S__doc__r keras.srcrkeras.src.api_exportrkeras.src.backendrrr&keras.src.backend.common.backend_utilsr keras.src.opsr keras.src.ops.operationr keras.src.ops.operation_utilsr r rr+r-r1r3r7r9r?rArErGrKrMrQrYrarcrgrirmrsrxrzr~rrrrrrrrrrrrrrrrrrr rrr&r.r4r7r<rCrIrOrTrdrfrjryrrrrrrr%rrrs DD------))))))222222555555*)))))------666666333339333!456676,33333I333 "678898,33333i333"$:;<<!!=<!.33333y333#%<=>>""?>"033333y333#%<=>>""?>".333339333433333333" %%  %. 3 3 3 3 3 3 3 3%'@ABBCCCCBC833333)333 # &&  &433333y333###: 3 3 3 3 3) 3 3 3 2344***54*6333339333!4566766 3 3 3 3 39 3 3 3!4566+++76+< 3 3 3 3 3i 3 3 3"$:;<<<0<0<0=<<0~ 3 3 3 3 3 3 3 3" ,4,4,4  ,4^=====i===D#%<=>>   0Q0Q0Q?>0Qf=====)===D #   222  2j"="="="="=9"="="=J!4566   444764n"="="="="=I"="="=J"%   999  9x(=(=(=(=(=I(=(=(=V"%  BBB  BJ,=,=,=,=,=I,=,=,=^"% BBB  BJJJJJJYJJJ>"$:;<<,,,=<,^========('* &&&  &RBBBBBiBBB6,/ 000  0fBBBBBIBBB<36 ...  .b'L'L'L'L'Ly'L'L'LT ;@$J$J$J  $JN     i   8 !L!L!L  !LH33333 3332'*?C222  2j<<<<