InstanceNormalization

Inheritance Diagram

---

class ashpy.layers.instance_normalization.InstanceNormalization(eps=1e-06, beta_initializer='zeros', gamma_initializer='ones')

Bases: tensorflow.python.keras.engine.base_layer.Layer

Instance Normalization Layer (used by Pix2Pix [1] and Pix2PixHD [2] ).

Basically it’s a normalization done at instance level. The implementation follows the basic implementation of the Batch Normalization Layer.

• Direct Usage:

  x = tf.ones((1, 10, 10, 64))
  
  # instantiate attention layer as model.
  normalization = InstanceNormalization()
  
  # evaluate passing x.
  output = normalization(x)
  
  # the output shape is.
  # the same as the input shape.
  print(output.shape)
  

  • Inside a Model:

    def MyModel():
        inputs = tf.keras.layers.Input(shape=[None, None, 64])
        normalization = InstanceNormalization()
        return tf.keras.Model(inputs=inputs, outputs=normalization(inputs))
    
    x = tf.ones((1, 10, 10, 64))
    model = MyModel()
    output = model(x)
    
    print(output.shape)
    

    (1, 10, 10, 64)
    

[1]	Image-to-Image Translation with Conditional Adversarial Networks https://arxiv.org/abs/1611.07004

[2]	High-Resolution Image Synthesis and Semantic Manipulation with Conditional GANs https://arxiv.org/abs/1711.11585

Methods

__init__([eps, beta_initializer, …])	Initialize the layer.
build(input_shape)	Assemble the layer.
call(inputs[, training])	Perform the computation.

Attributes

activity_regularizer	Optional regularizer function for the output of this layer.
dtype
dynamic
inbound_nodes	Deprecated, do NOT use! Only for compatibility with external Keras.
input	Retrieves the input tensor(s) of a layer.
input_mask	Retrieves the input mask tensor(s) of a layer.
input_shape	Retrieves the input shape(s) of a layer.
input_spec
losses	Losses which are associated with this Layer.
metrics
name	Returns the name of this module as passed or determined in the ctor.
name_scope	Returns a tf.name_scope instance for this class.
non_trainable_variables
non_trainable_weights
outbound_nodes	Deprecated, do NOT use! Only for compatibility with external Keras.
output	Retrieves the output tensor(s) of a layer.
output_mask	Retrieves the output mask tensor(s) of a layer.
output_shape	Retrieves the output shape(s) of a layer.
submodules	Sequence of all sub-modules.
trainable
trainable_variables	Sequence of variables owned by this module and it’s submodules.
trainable_weights
updates
variables	Returns the list of all layer variables/weights.
weights	Returns the list of all layer variables/weights.

__init__(eps=1e-06, beta_initializer='zeros', gamma_initializer='ones')

Initialize the layer.

Parameters:

• eps (float) – Variance_epsilon used by batch_norm layer.
• beta_initializer (str) – Initializer for the beta variable.
• gamma_initializer (str) – Initializer for the gamma variable.

Return type:

None

build(input_shape)

Assemble the layer.

Parameters:input_shape (tuple of (int)) – Specifies the shape of the input accepted by the layer. Return type:None

call(inputs, training=False)

Perform the computation.

Parameters:

• inputs (tf.Tensor) – Inputs for the computation.
• training (bool) – Controls for training or evaluation mode.

Return type:

Tensor

Returns:

tf.Tensor – Output Tensor.