B )`yf@sjdZddlmZddlmZddlmZddlmZddlmZddlm Z ddl m Z dd l m Z e d Gd d d eZe d dgdGddde jeZe ddgdGddde jeZe ddgdGdddeZe ddgdGddde jeZe ddgdGd d!d!e jeZe d"d#gdGd$d%d%e jeZe d&d'gdGd(d)d)e jeZe d*d+gdGd,d-d-e jeZe d.d/gdGd0d1d1e jeZe d2d3gdGd4d5d5eZe d6d7gdGd8d9d9eZe d:d;gdGdd?gdGd@dAdAeZe dBdCgdGdDdEdEeZe dFdGgdGdHdIdIeZdJdKZdLS)MzKeras initializers for TF 2. )absolute_import)division)print_function) constant_op)dtypes)backend) init_ops_v2) keras_exportzkeras.initializers.Initializerc@s.eZdZdZd ddZddZeddZdS) InitializeraInitializer base class: all Keras initializers inherit from this class. Initializers should implement a `__call__` method with the following signature: ```python def __call__(self, shape, dtype=None)`: # returns a tensor of shape `shape` and dtype `dtype` # containing values drawn from a distribution of your choice. ``` Optionally, you an also implement the method `get_config` and the class method `from_config` in order to support serialization -- just like with any Keras object. Here's a simple example: a random normal initializer. ```python import tensorflow as tf class ExampleRandomNormal(tf.keras.initializers.Initializer): def __init__(self, mean, stddev): self.mean = mean self.stddev = stddev def __call__(self, shape, dtype=None)`: return tf.random.normal( shape, mean=self.mean, stddev=self.stddev, dtype=dtype) def get_config(self): # To support serialization return {"mean": self.mean, "stddev": self.stddev} ``` Note that we don't have to implement `from_config` in the example above since the constructor arguments of the class the keys in the config returned by `get_config` are the same. In this case, the default `from_config` works fine. NcCstdS)zReturns a tensor object initialized as specified by the initializer. Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. N)NotImplementedError)selfshapedtyperc/home/dcms/DCMS/lib/python3.7/site-packages/tensorflow/python/keras/initializers/initializers_v2.py__call__GszInitializer.__call__cCsiS)zReturns the configuration of the initializer as a JSON-serializable dict. Returns: A JSON-serializable Python dict. r)r rrr get_configPszInitializer.get_configcCs|dd|f|S)auInstantiates an initializer from a configuration dictionary. Example: ```python initializer = RandomUniform(-1, 1) config = initializer.get_config() initializer = RandomUniform.from_config(config) ``` Args: config: A Python dictionary, the output of `get_config`. Returns: A `tf.keras.initializers.Initializer` instance. rN)pop)clsconfigrrr from_configXs zInitializer.from_config)N)__name__ __module__ __qualname____doc__rr classmethodrrrrrr s( r zkeras.initializers.Zeroszkeras.initializers.zeros)Zv1cs"eZdZdZdfdd ZZS)ZerosaInitializer that generates tensors initialized to 0. Also available via the shortcut function `tf.keras.initializers.zeros`. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.Zeros() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.Zeros() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Ncstt|j|t|dS)aReturns a tensor object initialized as specified by the initializer. Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. Only numeric or boolean dtypes are supported. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`). )r)superrr _get_dtype)r r r) __class__rrrs zZeros.__call__)N)rrrrr __classcell__rr)rrrnsrzkeras.initializers.Oneszkeras.initializers.onescs"eZdZdZdfdd ZZS)OnesaInitializer that generates tensors initialized to 1. Also available via the shortcut function `tf.keras.initializers.ones`. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.Ones() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.Ones() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Ncstt|j|t|dS)aReturns a tensor object initialized as specified by the initializer. Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. Only numeric or boolean dtypes are supported. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`). )r)rr!rr)r r r)rrrrs z Ones.__call__)N)rrrrrr rr)rrr!sr!zkeras.initializers.Constantzkeras.initializers.constantc@s,eZdZdZd ddZd ddZdd ZdS) ConstantaaInitializer that generates tensors with constant values. Also available via the shortcut function `tf.keras.initializers.constant`. Only scalar values are allowed. The constant value provided must be convertible to the dtype requested when calling the initializer. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.Constant(3.) >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.Constant(3.) >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: value: A Python scalar. rcCs ||_dS)N)value)r r#rrr__init__szConstant.__init__NcCstj|jt||dS)aMReturns a tensor object initialized to `self.value`. Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`). )rr )rZconstantr#r)r r rrrrrs zConstant.__call__cCs d|jiS)Nr#)r#)r rrrrszConstant.get_config)r)N)rrrrr$rrrrrrr"s  r"z keras.initializers.RandomUniformz!keras.initializers.random_uniformcs"eZdZdZdfdd ZZS) RandomUniforma{Initializer that generates tensors with a uniform distribution. Also available via the shortcut function `tf.keras.initializers.random_uniform`. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.RandomUniform(minval=0., maxval=1.) >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.RandomUniform(minval=0., maxval=1.) >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: minval: A python scalar or a scalar tensor. Lower bound of the range of random values to generate (inclusive). maxval: A python scalar or a scalar tensor. Upper bound of the range of random values to generate (exclusive). seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. Ncstt|j|t|dS)aReturns a tensor object initialized as specified by the initializer. Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. Only floating point and integer types are supported. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`). )r)rr%rr)r r r)rrrrs zRandomUniform.__call__)N)rrrrrr rr)rrr%sr%zkeras.initializers.RandomNormalz keras.initializers.random_normalcs"eZdZdZdfdd ZZS) RandomNormalaGInitializer that generates tensors with a normal distribution. Also available via the shortcut function `tf.keras.initializers.random_normal`. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.RandomNormal(mean=0., stddev=1.) >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.RandomNormal(mean=0., stddev=1.) >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: mean: a python scalar or a scalar tensor. Mean of the random values to generate. stddev: a python scalar or a scalar tensor. Standard deviation of the random values to generate. seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. Ncstt|j|t|dS)a}Returns a tensor object initialized to random normal values. Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. Only floating point types are supported. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`) )r)rr&rr)r r r)rrrrs zRandomNormal.__call__)N)rrrrrr rr)rrr&sr&z"keras.initializers.TruncatedNormalz#keras.initializers.truncated_normalcs"eZdZdZdfdd ZZS)TruncatedNormalaInitializer that generates a truncated normal distribution. Also available via the shortcut function `tf.keras.initializers.truncated_normal`. The values generated are similar to values from a `tf.keras.initializers.RandomNormal` initializer except that values more than two standard deviations from the mean are discarded and re-drawn. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.TruncatedNormal(mean=0., stddev=1.) >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.TruncatedNormal(mean=0., stddev=1.) >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: mean: a python scalar or a scalar tensor. Mean of the random values to generate. stddev: a python scalar or a scalar tensor. Standard deviation of the random values to generate. seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. Ncstt|j|t|dS)aReturns a tensor object initialized to random normal values (truncated). Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. Only floating point types are supported. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`) )r)rr'rr)r r r)rrrrLs zTruncatedNormal.__call__)N)rrrrrr rr)rrr'+sr'z"keras.initializers.VarianceScalingz#keras.initializers.variance_scalingcs"eZdZdZdfdd ZZS)VarianceScalingaInitializer capable of adapting its scale to the shape of weights tensors. Also available via the shortcut function `tf.keras.initializers.variance_scaling`. With `distribution="truncated_normal" or "untruncated_normal"`, samples are drawn from a truncated/untruncated normal distribution with a mean of zero and a standard deviation (after truncation, if used) `stddev = sqrt(scale / n)`, where `n` is: - number of input units in the weight tensor, if `mode="fan_in"` - number of output units, if `mode="fan_out"` - average of the numbers of input and output units, if `mode="fan_avg"` With `distribution="uniform"`, samples are drawn from a uniform distribution within `[-limit, limit]`, where `limit = sqrt(3 * scale / n)`. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.VarianceScaling( ... scale=0.1, mode='fan_in', distribution='uniform') >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.VarianceScaling( ... scale=0.1, mode='fan_in', distribution='uniform') >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: scale: Scaling factor (positive float). mode: One of "fan_in", "fan_out", "fan_avg". distribution: Random distribution to use. One of "truncated_normal", "untruncated_normal" and "uniform". seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. Ncstt|j|t|dS)aReturns a tensor object initialized as specified by the initializer. Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. Only floating point types are supported. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`) )r)rr(rr)r r r)rrrrs zVarianceScaling.__call__)N)rrrrrr rr)rrr(Ys(r(zkeras.initializers.Orthogonalzkeras.initializers.orthogonalcs"eZdZdZdfdd ZZS) OrthogonalaInitializer that generates an orthogonal matrix. Also available via the shortcut function `tf.keras.initializers.orthogonal`. If the shape of the tensor to initialize is two-dimensional, it is initialized with an orthogonal matrix obtained from the QR decomposition of a matrix of random numbers drawn from a normal distribution. If the matrix has fewer rows than columns then the output will have orthogonal rows. Otherwise, the output will have orthogonal columns. If the shape of the tensor to initialize is more than two-dimensional, a matrix of shape `(shape[0] * ... * shape[n - 2], shape[n - 1])` is initialized, where `n` is the length of the shape vector. The matrix is subsequently reshaped to give a tensor of the desired shape. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.Orthogonal() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.Orthogonal() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: gain: multiplicative factor to apply to the orthogonal matrix seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. References: [Saxe et al., 2014](https://openreview.net/forum?id=_wzZwKpTDF_9C) ([pdf](https://arxiv.org/pdf/1312.6120.pdf)) Ncstt|j|t|dS)a~Returns a tensor object initialized to an orthogonal matrix. Args: shape: Shape of the tensor. dtype: Optional dtype of the tensor. Only floating point types are supported. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`) )r)rr)rr)r r r)rrrrs zOrthogonal.__call__)N)rrrrrr rr)rrr)s%r)zkeras.initializers.Identityzkeras.initializers.identitycs"eZdZdZdfdd ZZS)IdentityaInitializer that generates the identity matrix. Also available via the shortcut function `tf.keras.initializers.identity`. Only usable for generating 2D matrices. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.Identity() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.Identity() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: gain: Multiplicative factor to apply to the identity matrix. Ncstt|j|t|dS)aReturns a tensor object initialized to a 2D identity matrix. Args: shape: Shape of the tensor. It should have exactly rank 2. dtype: Optional dtype of the tensor. Only floating point types are supported. If not specified, `tf.keras.backend.floatx()` is used, which default to `float32` unless you configured it otherwise (via `tf.keras.backend.set_floatx(float_dtype)`) )r)rr*rr)r r r)rrrrs zIdentity.__call__)N)rrrrrr rr)rrr*sr*z keras.initializers.GlorotUniformz!keras.initializers.glorot_uniformcs*eZdZdZdfdd ZddZZS) GlorotUniformaThe Glorot uniform initializer, also called Xavier uniform initializer. Also available via the shortcut function `tf.keras.initializers.glorot_uniform`. Draws samples from a uniform distribution within `[-limit, limit]`, where `limit = sqrt(6 / (fan_in + fan_out))` (`fan_in` is the number of input units in the weight tensor and `fan_out` is the number of output units). Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.GlorotUniform() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.GlorotUniform() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. References: [Glorot et al., 2010](http://proceedings.mlr.press/v9/glorot10a.html) ([pdf](http://jmlr.org/proceedings/papers/v9/glorot10a/glorot10a.pdf)) Ncstt|jddd|ddS)Ng?fan_avguniform)scalemode distributionseed)rr+r$)r r1)rrrr$ s  zGlorotUniform.__init__cCs d|jiS)Nr1)r1)r rrrrszGlorotUniform.get_config)N)rrrrr$rr rr)rrr+sr+zkeras.initializers.GlorotNormalz keras.initializers.glorot_normalcs*eZdZdZdfdd ZddZZS) GlorotNormalaThe Glorot normal initializer, also called Xavier normal initializer. Also available via the shortcut function `tf.keras.initializers.glorot_normal`. Draws samples from a truncated normal distribution centered on 0 with `stddev = sqrt(2 / (fan_in + fan_out))` where `fan_in` is the number of input units in the weight tensor and `fan_out` is the number of output units in the weight tensor. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.GlorotNormal() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.GlorotNormal() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Args: seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. References: [Glorot et al., 2010](http://proceedings.mlr.press/v9/glorot10a.html) ([pdf](http://jmlr.org/proceedings/papers/v9/glorot10a/glorot10a.pdf)) Ncstt|jddd|ddS)Ng?r,truncated_normal)r.r/r0r1)rr2r$)r r1)rrrr$5s  zGlorotNormal.__init__cCs d|jiS)Nr1)r1)r rrrr<szGlorotNormal.get_config)N)rrrrr$rr rr)rrr2sr2zkeras.initializers.LecunNormalzkeras.initializers.lecun_normalcs*eZdZdZdfdd ZddZZS) LecunNormalaLecun normal initializer. Also available via the shortcut function `tf.keras.initializers.lecun_normal`. Initializers allow you to pre-specify an initialization strategy, encoded in the Initializer object, without knowing the shape and dtype of the variable being initialized. Draws samples from a truncated normal distribution centered on 0 with `stddev = sqrt(1 / fan_in)` where `fan_in` is the number of input units in the weight tensor. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.LecunNormal() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.LecunNormal() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Arguments: seed: A Python integer. Used to seed the random generator. References: - Self-Normalizing Neural Networks, [Klambauer et al., 2017] (https://papers.nips.cc/paper/6698-self-normalizing-neural-networks) ([pdf] (https://papers.nips.cc/paper/6698-self-normalizing-neural-networks.pdf)) - Efficient Backprop, [Lecun et al., 1998](http://yann.lecun.com/exdb/publis/pdf/lecun-98b.pdf) Ncstt|jddd|ddS)Ng?fan_inr3)r.r/r0r1)rr4r$)r r1)rrrr$hs zLecunNormal.__init__cCs d|jiS)Nr1)r1)r rrrrlszLecunNormal.get_config)N)rrrrr$rr rr)rrr4@s&r4zkeras.initializers.LecunUniformz keras.initializers.lecun_uniformcs*eZdZdZdfdd ZddZZS) LecunUniformagLecun uniform initializer. Also available via the shortcut function `tf.keras.initializers.lecun_uniform`. Draws samples from a uniform distribution within `[-limit, limit]`, where `limit = sqrt(3 / fan_in)` (`fan_in` is the number of input units in the weight tensor). Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.LecunUniform() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.LecunUniform() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Arguments: seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. References: - Self-Normalizing Neural Networks, [Klambauer et al., 2017](https://papers.nips.cc/paper/6698-self-normalizing-neural-networks) # pylint: disable=line-too-long ([pdf](https://papers.nips.cc/paper/6698-self-normalizing-neural-networks.pdf)) - Efficient Backprop, [Lecun et al., 1998](http://yann.lecun.com/exdb/publis/pdf/lecun-98b.pdf) Ncstt|jddd|ddS)Ng?r5r-)r.r/r0r1)rr6r$)r r1)rrrr$s zLecunUniform.__init__cCs d|jiS)Nr1)r1)r rrrrszLecunUniform.get_config)N)rrrrr$rr rr)rrr6ps!r6zkeras.initializers.HeNormalzkeras.initializers.he_normalcs*eZdZdZdfdd ZddZZS)HeNormala He normal initializer. Also available via the shortcut function `tf.keras.initializers.he_normal`. It draws samples from a truncated normal distribution centered on 0 with `stddev = sqrt(2 / fan_in)` where `fan_in` is the number of input units in the weight tensor. Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.HeNormal() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.HeNormal() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Arguments: seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. References: [He et al., 2015](https://www.cv-foundation.org/openaccess/content_iccv_2015/html/He_Delving_Deep_into_ICCV_2015_paper.html) # pylint: disable=line-too-long ([pdf](https://www.cv-foundation.org/openaccess/content_iccv_2015/papers/He_Delving_Deep_into_ICCV_2015_paper.pdf)) Ncstt|jddd|ddS)Ng@r5r3)r.r/r0r1)rr7r$)r r1)rrrr$s zHeNormal.__init__cCs d|jiS)Nr1)r1)r rrrrszHeNormal.get_config)N)rrrrr$rr rr)rrr7sr7zkeras.initializers.HeUniformzkeras.initializers.he_uniformcs*eZdZdZdfdd ZddZZS) HeUniformaHe uniform variance scaling initializer. Also available via the shortcut function `tf.keras.initializers.he_uniform`. Draws samples from a uniform distribution within `[-limit, limit]`, where `limit = sqrt(6 / fan_in)` (`fan_in` is the number of input units in the weight tensor). Examples: >>> # Standalone usage: >>> initializer = tf.keras.initializers.HeUniform() >>> values = initializer(shape=(2, 2)) >>> # Usage in a Keras layer: >>> initializer = tf.keras.initializers.HeUniform() >>> layer = tf.keras.layers.Dense(3, kernel_initializer=initializer) Arguments: seed: A Python integer. An initializer created with a given seed will always produce the same random tensor for a given shape and dtype. References: [He et al., 2015](https://www.cv-foundation.org/openaccess/content_iccv_2015/html/He_Delving_Deep_into_ICCV_2015_paper.html) # pylint: disable=line-too-long ([pdf](https://www.cv-foundation.org/openaccess/content_iccv_2015/papers/He_Delving_Deep_into_ICCV_2015_paper.pdf)) Ncstt|jddd|ddS)Ng@r5r-)r.r/r0r1)rr8r$)r r1)rrrr$s zHeUniform.__init__cCs d|jiS)Nr1)r1)r rrrrszHeUniform.get_config)N)rrrrr$rr rr)rrr8sr8cCs|dkrt}t|S)N)rZfloatxrZas_dtype)rrrrrsrN) r __future__rrrZtensorflow.python.frameworkrrZtensorflow.python.kerasrZtensorflow.python.opsrZ tensorflow.python.util.tf_exportr objectr rr!r"r%r&r'r(r)r*r+r2r4r6r7r8rrrrrs        P  + ' & + 4 1 "()-(%%