o 2h,@s dZddlmZddlmZddlmZddlmZddlm Z ddlm Z ddlm Z dd lm Z dd l mZGd d d ZGd ddeZGdddeZGdddeZGdddeZGdddeZeZeZeZeZeZeZeZeZddZdddZddZdS)z@Constraints: functions that impose constraints on weight values.) tensor_shape)backend)deserialize_keras_objectserialize_keras_object) array_ops)array_ops_stack)math_ops) while_loop) doc_controlsc@s eZdZdZddZddZdS) ConstraintaKBase 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()) cCs|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). selfwr r ^/var/www/html/chatgem/venv/lib/python3.10/site-packages/tensorflow/python/keras/constraints.py__call__6s zConstraint.__call__cCsiS)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 rr r r get_configEs zConstraint.get_configN)__name__ __module__ __qualname____doc__rrr r r rr s r c@s6eZdZdZd ddZejddZejdd Zd S) MaxNorma/MaxNorm 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)`. rcCs||_||_dSN max_valueaxis)rrrr r r__init__is zMaxNorm.__init__cCsBttjt||jdd}t|d|j}||t|S)NTrkeepdimsr) rsqrtr reduce_sumsquarercliprepsilonrrnormsdesiredr r rrms zMaxNorm.__call__cCs|j|jdS)Nrrrr r rrtszMaxNorm.get_configN)rr rrrrrr do_not_generate_docsrrr r r rrQs  rc@seZdZdZddZdS)NonNegzyConstrains the weights to be non-negative. Also available via the shortcut function `tf.keras.constraints.non_neg`. cCs|tt|dtS)N)r cast greater_equalrfloatxrr r rrszNonNeg.__call__N)rrrrrr r r rr,ys r,c@s6eZdZdZd ddZejddZejddZd S) 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)`. rcCs ||_dSrr)rrr r rrs zUnitNorm.__init__c Cs*|tttjt||jddS)NTr )rr&r"r r#r$rrr r rrs zUnitNorm.__call__cCs d|jiS)Nrr2rr r rrs zUnitNorm.get_configN)rr*r r r rr1s  r1c@s6eZdZdZd ddZejddZejd d Zd S) MinMaxNormaMinMaxNorm weight constraint. Constrains the weights incident to each hidden unit to have the norm between a lower bound and an upper bound. Also available via the shortcut function `tf.keras.constraints.min_max_norm`. Args: min_value: the minimum norm for the incoming weights. max_value: the maximum norm for the incoming weights. rate: rate for enforcing the constraint: weights will be rescaled to yield `(1 - rate) * norm + rate * norm.clip(min_value, max_value)`. Effectively, this means that rate=1.0 stands for strict enforcement of the constraint, while rate<1.0 means that weights will be rescaled at each step to slowly move towards a value inside the desired interval. axis: integer, axis along which to calculate weight norms. For instance, in a `Dense` layer the weight matrix has shape `(input_dim, output_dim)`, set `axis` to `0` to constrain each weight vector of length `(input_dim,)`. In a `Conv2D` layer with `data_format="channels_last"`, the weight tensor has shape `(rows, cols, input_depth, output_depth)`, set `axis` to `[0, 1, 2]` to constrain the weights of each filter tensor of size `(rows, cols, input_depth)`. r-?rcCs||_||_||_||_dSr min_valuerrater)rr6rr7rr r rrs zMinMaxNorm.__init__cCsXttjt||jdd}|jt||j|j d|j|}||t |S)NTr ) rr"r r#r$rr7r%r6rr&r'r r rrs zMinMaxNorm.__call__cCs|j|j|j|jdS)Nr5r5rr r rrs zMinMaxNorm.get_configN)r-r4r4rr*r r r rr3s  r3c@s&eZdZdZejddZddZdS)RadialConstraintasConstrains `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)`. cCs|j}|jdus |jdkrtd||\}}}}t|||||f}t|jtjtj |dddd}ttjtj |dddd||||fS)Nz8The weight tensor must be of rank 4, but is of shape: %sr2r) shaperank ValueErrorrreshapemap_fn_kernel_constraintstackrunstack)rrw_shapeheightwidthchannelskernelsr r rrs  zRadialConstraint.__call__c stjddgddggddtd}t|ddttt|ddfddfd d}ttt|ddd dd d}fd d}fd d}tj||||g|t ddggd\}}|S)zCRadially constraints a kernel with shape (height, width, channels).r8int32dtyperrboolcsddfS)Nr8r r kernelstartr rsz5RadialConstraint._kernel_constraint..cs,ddftjdjdS)Nr8)rrrJ)rzerosrKr rMr rrPscStjdddS)NrrIrJrconstantr r r rrPcSrR)Nr8rIrJrSr r r rrPrUcs t|Sr)rless)indexargs)rOr rrPs cs(|dtj|||fdfS)Nr8)constant_values)rpad)iarrayrNpaddingrOr rbody_fns z4RadialConstraint._kernel_constraint..body_fnN)shape_invariants) rrTr<r.switchr floormodr get_shaper TensorShape)rrN kernel_shape kernel_newrWwhile_conditionr__r r]rrA s.     z#RadialConstraint._kernel_constraintN)rrrrr r+rrAr r r rr9s   r9cCst|Srr) constraintr r r serialize=srjNcCst|t|ddS)Nri)module_objectscustom_objectsprintable_module_name)rglobals)configrlr r r deserializeAs rpcCsZ|durdSt|trt|St|trt|id}t|St|r%|Stdt|)N) class_nameroz+Could not interpret constraint identifier: ) isinstancedictrpstrcallabler>) identifierror r rgetIs  rwr) rtensorflow.python.frameworkrtensorflow.python.kerasr+tensorflow.python.keras.utils.generic_utilsrrtensorflow.python.opsrrr r tensorflow.tools.docsr r rr,r1r3r9max_normnon_neg unit_norm min_max_normradial_constraintmaxnormnonnegunitnormrjrprwr r r rs6         3( "8T