~Wh4|dZddlmcmZddlmZddlmZ ddl m Z ddl m Z ddl mZddlmZed Gd d Zed d GddeZeddGddeZeddGddeZeddGddeZeddGddeZeZeZeZeZeZeZeZeZed d'd"Zed#d(d$Z ed%d&Z!dS))z@Constraints: functions that impose constraints on weight values.N)backend) serialization)deserialize_keras_object)serialize_keras_object) keras_export) doc_controlszkeras.constraints.Constraintc4eZdZdZdZdZedZdS) ConstraintaoBase class for weight constraints. A `Constraint` instance works like a stateless function. Users who subclass this class should override the `__call__` method, which takes a single weight parameter and return a projected version of that parameter (e.g. normalized or clipped). Constraints can be used with various Keras layers via the `kernel_constraint` or `bias_constraint` arguments. Here's a simple example of a non-negative weight constraint: >>> class NonNegative(tf.keras.constraints.Constraint): ... ... def __call__(self, w): ... return w * tf.cast(tf.math.greater_equal(w, 0.), w.dtype) >>> weight = tf.constant((-1.0, 1.0)) >>> NonNegative()(weight) >>> tf.keras.layers.Dense(4, kernel_constraint=NonNegative()) c|S)agApplies the constraint to the input weight variable. By default, the inputs weight variable is not modified. Users should override this method to implement their own projection function. Args: w: Input weight variable. Returns: Projected variable (by default, returns unmodified inputs). selfws W/var/www/html/movieo_spanner_bot/venv/lib/python3.11/site-packages/keras/constraints.py__call__zConstraint.__call__9s ciS)aReturns a Python dict of the object config. A constraint config is a Python dictionary (JSON-serializable) that can be used to reinstantiate the same object. Returns: Python dict containing the configuration of the constraint object. r rs r get_configzConstraint.get_configHs  rc|di|S)aInstantiates a weight constraint from a configuration dictionary. Example: ```python constraint = UnitNorm() config = constraint.get_config() constraint = UnitNorm.from_config(config) ``` Args: config: A Python dictionary, the output of `get_config`. Returns: A `tf.keras.constraints.Constraint` instance. r r )clsconfigs r from_configzConstraint.from_configSs$s}}V}}rN)__name__ __module__ __qualname____doc__rr classmethodrr rrr r sW0      [rr zkeras.constraints.MaxNormzkeras.constraints.max_normcZeZdZdZddZejdZejdZdS) MaxNormaQMaxNorm weight constraint. Constrains the weights incident to each hidden unit to have a norm less than or equal to a desired value. Also available via the shortcut function `tf.keras.constraints.max_norm`. Args: max_value: the maximum norm value for the incoming weights. axis: integer, axis along which to calculate weight norms. For instance, in a `Dense` layer the weight matrix has shape `(input_dim, output_dim)`, set `axis` to `0` to constrain each weight vector of length `(input_dim,)`. In a `Conv2D` layer with `data_format="channels_last"`, the weight tensor has shape `(rows, cols, input_depth, output_depth)`, set `axis` to `[0, 1, 2]` to constrain the weights of each filter tensor of size `(rows, cols, input_depth)`. rc"||_||_dSN max_valueaxis)rr%r&s r__init__zMaxNorm.__init__s" rctjtjtj||jd}tj|d|j}||tj|zz zS)NTr&keepdimsr) rsqrttf reduce_sumsquarer&clipr%epsilonrrnormsdesireds rrzMaxNorm.__call__se M")A,,TY F F F  ,ua88Gw00589::rc |j|jdS)Nr$r$rs rrzMaxNorm.get_configs!^TY???rN)r!r rrrrr'rdo_not_generate_docsrrr rrr r hsu.&;;'&;&@@'&@@@rr zkeras.constraints.NonNegzkeras.constraints.non_negceZdZdZdZdS)NonNegz}Constrains the weights to be non-negative. Also available via the shortcut function `tf.keras.constraints.non_neg`. cz|tjtj|dtjzS)N)r,cast greater_equalrfloatxr s rrzNonNeg.__call__s.272+As33W^5E5EFFFFrN)rrrrrr rrr8r8s2 GGGGGrr8zkeras.constraints.UnitNormzkeras.constraints.unit_normcZeZdZdZddZejdZejdZdS)UnitNormaConstrains the weights incident to each hidden unit to have unit norm. Also available via the shortcut function `tf.keras.constraints.unit_norm`. Args: axis: integer, axis along which to calculate weight norms. For instance, in a `Dense` layer the weight matrix has shape `(input_dim, output_dim)`, set `axis` to `0` to constrain each weight vector of length `(input_dim,)`. In a `Conv2D` layer with `data_format="channels_last"`, the weight tensor has shape `(rows, cols, input_depth, output_depth)`, set `axis` to `[0, 1, 2]` to constrain the weights of each filter tensor of size `(rows, cols, input_depth)`. rc||_dSr#r&)rr&s rr'zUnitNorm.__init__s  rc |tjtjtjtj||jdzz S)NTr))rr0r+r,r-r.r&r s rrzUnitNorm.__call__sL O  l billTJJJ   rcd|jiS)Nr&rArs rrzUnitNorm.get_configs ""rN)rr5r rrr?r?sp$&  '& &##'&###rr?zkeras.constraints.MinMaxNormzkeras.constraints.min_max_normcZeZdZdZd dZejdZejdZdS) MinMaxNorma(MinMaxNorm weight constraint. Constrains the weights incident to each hidden unit to have the norm between a lower bound and an upper bound. Also available via the shortcut function `tf.keras.constraints.min_max_norm`. Args: min_value: the minimum norm for the incoming weights. max_value: the maximum norm for the incoming weights. rate: rate for enforcing the constraint: weights will be rescaled to yield `(1 - rate) * norm + rate * norm.clip(min_value, max_value)`. Effectively, this means that rate=1.0 stands for strict enforcement of the constraint, while rate<1.0 means that weights will be rescaled at each step to slowly move towards a value inside the desired interval. axis: integer, axis along which to calculate weight norms. For instance, in a `Dense` layer the weight matrix has shape `(input_dim, output_dim)`, set `axis` to `0` to constrain each weight vector of length `(input_dim,)`. In a `Conv2D` layer with `data_format="channels_last"`, the weight tensor has shape `(rows, cols, input_depth, output_depth)`, set `axis` to `[0, 1, 2]` to constrain the weights of each filter tensor of size `(rows, cols, input_depth)`. r:?rc>||_||_||_||_dSr# min_valuer%rater&)rrIr%rJr&s rr'zMinMaxNorm.__init__s"""  rc&tjtjtj||jd}|jtj||j|j zd|jz |zz}||tj |zz zS)NTr)) rr+r,r-r.r&rJr/rIr%r0r1s rrzMinMaxNorm.__call__s M")A,,TY F F F   I UDNDNKK K49}% & Gw00589::rc8|j|j|j|jdS)NrHrHrs rrzMinMaxNorm.get_configs'II    rN)r:rFrFrr5r rrrErEsp> &;;'&;&  '&   rrEz"keras.constraints.RadialConstraintz#keras.constraints.radial_constraintc8eZdZdZejdZdZdS)RadialConstraintaConstrains `Conv2D` kernel weights to be the same for each radius. Also available via the shortcut function `tf.keras.constraints.radial_constraint`. For example, the desired output for the following 4-by-4 kernel: ``` kernel = [[v_00, v_01, v_02, v_03], [v_10, v_11, v_12, v_13], [v_20, v_21, v_22, v_23], [v_30, v_31, v_32, v_33]] ``` is this:: ``` kernel = [[v_11, v_11, v_11, v_11], [v_11, v_33, v_33, v_11], [v_11, v_33, v_33, v_11], [v_11, v_11, v_11, v_11]] ``` This constraint can be applied to any `Conv2D` layer version, including `Conv2DTranspose` and `SeparableConv2D`, and with either `"channels_last"` or `"channels_first"` data format. The method assumes the weight tensor is of shape `(rows, cols, input_depth, output_depth)`. c |j}|j |jdkrtd||\}}}}tj|||||zf}tj|jtjtj |dd}tjtjtj |dd||||fS)NzGThe weight tensor must have rank 4. Received weight tensor with shape: rAr) shaperank ValueErrorrreshapemap_fn_kernel_constraintstackr,unstack)rrw_shapeheightwidthchannelskernelss rrzRadialConstraint.__call__s' < 7<1#4#4@6=@@  ,3(x OAx'/AB C C N  # M"*QR000q 9 9 9   M"*QQ///b 9 9 9 UHg .   rc  tjddgddggdtjd}tj|dz d tjtjt j|dd fd fd}tjtjt j|ddd d } fd } fd }t jj ||||g| t j d d gg\}}|S)zKRadially constraints a kernel with shape (height, width, channels).rLint32dtyperr!boolc,dz dz fS)NrLr kernelstartsrz5RadialConstraint._kernel_constraint..?s#F519u,eai%.??@rcddz dz ftjdjzS)NrL)r!r!rb)rzerosrcrfsrriz5RadialConstraint._kernel_constraint..@s<F519u,eai%.??@mF&,7778rc.tjddS)Nrrarbrconstantr rrriz5RadialConstraint._kernel_constraint..EG$Qg666rc.tjddS)NrLrarbrmr rrriz5RadialConstraint._kernel_constraint..Frorc.tj|Sr#)rless)indexargsrhs rriz5RadialConstraint._kernel_constraint..Hsw|E5/I/Irc X|dztj||z|zffS)NrL)constant_values)r,pad)iarrayrgpaddingrhs rbody_fnz4RadialConstraint._kernel_constraint..body_fnJs@q5"&wuqy%!)7K0L rN)shape_invariants) rrnrSr;switchr,mathfloormodcompatv1 while_loop get_shape TensorShape) rrg kernel_shape kernel_newrswhile_conditionr{_rzrhs ` @@rrXz#RadialConstraint._kernel_constraint5s{"QFQF#37CCC}V,,Q/  \A-w77^ L)),::F C C @ @ @ @ @ 8 8 8 8 8   L)),::F C C 6 6 6 6   JIII         //   J #oo//t 1M1MN 0  : rN)rrrrrr6rrXr rrrOrOsL:&  '& *     rrOzkeras.constraints.serializeFcL|rtj|St|Sr#)legacy_serializationr) constraintuse_legacy_formats r serializerfs*G#::FFF !* - --rzkeras.constraints.deserializec|r$tj|t|dSt|t|dS)Nr)module_objectscustom_objectsprintable_module_name)rrglobals)rrrs r deserializerms] #< "99)".     $yy%*    rzkeras.constraints.getc|dSt|trd|v}t||St|tr!t|id}t|St |r|St d|)z&Retrieves a Keras constraint function.Nmodule)r) class_namerz4Could not interpret constraint function identifier: ) isinstancedictrstrcallablerU) identifierrrs rgetr~st*d##  $J6:9JKKKK J $ $ #J2>>6""" *    O: O O   r)F)NF)"rtensorflow.compat.v2rv2r,kerasrkeras.saving.legacyrr!keras.saving.legacy.serializationrr tensorflow.python.util.tf_exportrtensorflow.tools.docsrr r r8r?rErOmax_normnon_neg unit_norm min_max_normradial_constraintmaxnormnonnegunitnormrrrr rrrs"GF!!!!!!!!!EEEEEEFFFFFFDDDDDD:99999......,--EEEEEEE.-EP)+GHH&@&@&@&@&@j&@&@IH&@R(*EFFGGGGGZGGGFG*,IJJ!#!#!#!#!#z!#!#KJ!#H,.NOO8 8 8 8 8 8 8 PO8 v(*OTTTTTzTTTr     $   +,,...-,. -..   /.  %&&  '&   r