tf.keras.layers.Layer

TensorFlow 1 version

View source on GitHub

Base layer class.

Inherits From: Module

tf.keras.layers.Layer(
    trainable=True, name=None, dtype=None, dynamic=False, **kwargs
)

This is the class from which all layers inherit.

A layer is a class implementing common neural networks operations, such as convolution, batch norm, etc. These operations require managing weights, losses, updates, and inter-layer connectivity.

Users will just instantiate a layer and then treat it as a callable.

We recommend that descendants of Layer implement the following methods:

• __init__(): Save configuration in member variables
• build(): Called once from __call__, when we know the shapes of inputs and dtype. Should have the calls to add_weight(), and then call the super's build() (which sets self.built = True, which is nice in case the user wants to call build() manually before the first __call__).
• call(): Called in __call__ after making sure build() has been called once. Should actually perform the logic of applying the layer to the input tensors (which should be passed in as the first argument).

Read-only properties: name: The name of the layer (string). dtype: The dtype of the layer's computations and weights. If mixed precision is used with a tf.keras.mixed_precision.experimental.Policy, this is instead just the dtype of the layer's weights, as the computations are done in a different dtype. updates: List of update ops of this layer. losses: List of losses added by this layer. trainable_weights: List of variables to be included in backprop. non_trainable_weights: List of variables that should not be included in backprop. weights: The concatenation of the lists trainable_weights and non_trainable_weights (in this order).

Mutable properties:

• trainable: Whether the layer should be trained (boolean).
• input_spec: Optional (list of) InputSpec object(s) specifying the constraints on inputs that can be accepted by the layer.

Dtypes and casting

Each layer has a dtype, which is typically the dtype of the layer's computations and variables. A layer's dtype can be queried via the Layer.dtype property. The dtype is specified with the dtype constructor argument. In TensorFlow 2, the dtype defaults to tf.keras.backend.floatx() if no dtype is passed. floatx() itself defaults to "float32". Additionally, layers will cast their inputs to the layer's dtype in TensorFlow 2. For example:

x = tf.ones((4, 4, 4, 4), dtype='float64')
layer = tf.keras.layers.Conv2D(filters=4, kernel_size=2)
print(layer.dtype)  # float32

# `layer` casts it's inputs to layer.dtype, which is float32, and does
# computations in float32.
y = layer(x)

Currently, only tensors in the first argument to the layer's call method are casted. For example:

class MyLayer(tf.keras.layers.Layer):
  # Bug! `b` will not be casted.
  def call(self, a, b):
    return a + 1., b + 1.

a = tf.constant(1., dtype="float32")
b = tf.constant(1., dtype="float32")

layer = MyLayer(dtype="float64")
x, y = layer(a, b)
print(x.dtype)  # float64
print(y.dtype)  # float32. Not casted since `b` was not passed to first input

It is recommended to accept tensors only in the first argument. This way, all tensors are casted to the layer's dtype. MyLayer should therefore be written as:

class MyLayer(tf.keras.layers.Layer):
  # Now, all tensor inputs will be casted.
  def call(self, inputs):
    a, b = inputs
    return a + 1., b + 1.

a = tf.constant(1., dtype="float32")
b = tf.constant(1., dtype="float32")

layer = MyLayer(dtype="float64")
x, y = layer((a, b))
print(x.dtype)  # float64
print(y.dtype)  # float64.

In a future minor release, tensors in other arguments may be casted as well.

Currently, other arguments are not automatically casted for technical reasons, but this may change in a future minor release.

A layer subclass can prevent its inputs from being autocasted by passing autocast=False to the layer constructor. For example:

class MyLayer(tf.keras.layers.Layer):

  def __init__(self, **kwargs):
    kwargs['autocast']=False
    super(MyLayer, self).__init__(**kwargs)

  def call(self, inp):
    return inp

x = tf.ones((4, 4, 4, 4), dtype='float64')
layer = MyLayer()
print(layer.dtype)  # float32.
y = layer(x)  # MyLayer will not cast inputs to it's dtype of float32
print(y.dtype)  # float64

Running models in float64 in TensorFlow 2

If you want to run a Model in float64, you can set floatx to be float64 by calling tf.keras.backend.set_floatx('float64'). This will cause all layers to default to float64 instead of float32:

tf.keras.backend.set_floatx('float64')
layer1 = tf.keras.layers.Dense(4)
layer2 = tf.keras.layers.Dense(4)

x = tf.ones((4, 4))
y = layer2(layer1(x))  # Both layers run in float64

Alternatively, you can pass dtype='float64' to each individual layer. Note that if you have any layers which contain other layers as members, you must ensure each sublayer gets dtype='float64' passed to it's constructor as well:

layer1 = tf.keras.layers.Dense(4, dtype='float64')
layer2 = tf.keras.layers.Dense(4, dtype='float64')

x = tf.ones((4, 4))
y = layer2(layer1(x))  # Both layers run in float64

class NestedLayer(tf.keras.layers.Layer):
  def __init__(self, **kwargs):
    super(NestedLayer, self).__init__(**kwargs)
    self.dense = tf.keras.layers.Dense(4, dtype=kwargs.get('dtype'))

  def call(self, inp):
    return self.dense(inp)

layer3 = NestedLayer(dtype='float64')
z = layer3(x)  # layer3's dense layer runs in float64, since NestedLayer
               # correcty passed it's dtype to it's dense layer

Methods

add_loss

View source

add_loss(
    losses, inputs=None
)

Add loss tensor(s), potentially dependent on layer inputs.

Some losses (for instance, activity regularization losses) may be dependent on the inputs passed when calling a layer. Hence, when reusing the same layer on different inputs a and b, some entries in layer.losses may be dependent on a and some on b. This method automatically keeps track of dependencies.

This method can be used inside a subclassed layer or model's call function, in which case losses should be a Tensor or list of Tensors.

Example:

class MyLayer(tf.keras.layers.Layer):
  def call(inputs, self):
    self.add_loss(tf.abs(tf.reduce_mean(inputs)), inputs=True)
    return inputs

This method can also be called directly on a Functional Model during construction. In this case, any loss Tensors passed to this Model must be symbolic and be able to be traced back to the model's Inputs. These losses become part of the model's topology and are tracked in get_config.

Example:

inputs = tf.keras.Input(shape=(10,))
x = tf.keras.layers.Dense(10)(inputs)
outputs = tf.keras.layers.Dense(1)(x)
model = tf.keras.Model(inputs, outputs)
# Actvity regularization.
model.add_loss(tf.abs(tf.reduce_mean(x)))

If this is not the case for your loss (if, for example, your loss references a Variable of one of the model's layers), you can wrap your loss in a zero-argument lambda. These losses are not tracked as part of the model's topology since they can't be serialized.

Example:

inputs = tf.keras.Input(shape=(10,))
x = tf.keras.layers.Dense(10)(inputs)
outputs = tf.keras.layers.Dense(1)(x)
model = tf.keras.Model(inputs, outputs)
# Weight regularization.
model.add_loss(lambda: tf.reduce_mean(x.kernel))

The get_losses_for method allows to retrieve the losses relevant to a specific set of inputs.

add_metric

View source

add_metric(
    value, aggregation=None, name=None
)

Adds metric tensor to the layer.

add_update

View source

add_update(
    updates, inputs=None
)

Add update op(s), potentially dependent on layer inputs. (deprecated arguments)

Weight updates (for instance, the updates of the moving mean and variance in a BatchNormalization layer) may be dependent on the inputs passed when calling a layer. Hence, when reusing the same layer on different inputs a and b, some entries in layer.updates may be dependent on a and some on b. This method automatically keeps track of dependencies.

The get_updates_for method allows to retrieve the updates relevant to a specific set of inputs.

This call is ignored when eager execution is enabled (in that case, variable updates are run on the fly and thus do not need to be tracked for later execution).

add_weight

View source

add_weight(
    name=None, shape=None, dtype=None, initializer=None, regularizer=None,
    trainable=None, constraint=None, partitioner=None, use_resource=None,
    synchronization=tf.VariableSynchronization.AUTO,
    aggregation=tf.compat.v1.VariableAggregation.NONE, **kwargs
)

Adds a new variable to the layer.

build

View source

build(
    input_shape
)

Creates the variables of the layer (optional, for subclass implementers).

This is a method that implementers of subclasses of Layer or Model can override if they need a state-creation step in-between layer instantiation and layer call.

This is typically used to create the weights of Layer subclasses.

call

View source

call(
    inputs, **kwargs
)

This is where the layer's logic lives.

compute_mask

View source

compute_mask(
    inputs, mask=None
)

Computes an output mask tensor.

compute_output_shape

View source

compute_output_shape(
    input_shape
)

Computes the output shape of the layer.

If the layer has not been built, this method will call build on the layer. This assumes that the layer will later be used with inputs that match the input shape provided here.

compute_output_signature

View source

compute_output_signature(
    input_signature
)

Compute the output tensor signature of the layer based on the inputs.

Unlike a TensorShape object, a TensorSpec object contains both shape and dtype information for a tensor. This method allows layers to provide output dtype information if it is different from the input dtype. For any layer that doesn't implement this function, the framework will fall back to use compute_output_shape, and will assume that the output dtype matches the input dtype.

count_params

View source

count_params()

Count the total number of scalars composing the weights.

from_config

View source

@classmethod
from_config(
    config
)

Creates a layer from its config.

This method is the reverse of get_config, capable of instantiating the same layer from the config dictionary. It does not handle layer connectivity (handled by Network), nor weights (handled by set_weights).

get_config

View source

get_config()

Returns the config of the layer.

A layer config is a Python dictionary (serializable) containing the configuration of a layer. The same layer can be reinstantiated later (without its trained weights) from this configuration.

The config of a layer does not include connectivity information, nor the layer class name. These are handled by Network (one layer of abstraction above).

get_input_at

View source

get_input_at(
    node_index
)

Retrieves the input tensor(s) of a layer at a given node.

get_input_mask_at

View source

get_input_mask_at(
    node_index
)

Retrieves the input mask tensor(s) of a layer at a given node.

get_input_shape_at

View source

get_input_shape_at(
    node_index
)

Retrieves the input shape(s) of a layer at a given node.

get_losses_for

View source

get_losses_for(
    inputs
)

Retrieves losses relevant to a specific set of inputs.

get_output_at

View source

get_output_at(
    node_index
)

Retrieves the output tensor(s) of a layer at a given node.

get_output_mask_at

View source

get_output_mask_at(
    node_index
)

Retrieves the output mask tensor(s) of a layer at a given node.

get_output_shape_at

View source

get_output_shape_at(
    node_index
)

Retrieves the output shape(s) of a layer at a given node.

get_updates_for

View source

get_updates_for(
    inputs
)

Retrieves updates relevant to a specific set of inputs.

get_weights

View source

get_weights()

Returns the current weights of the layer.

set_weights

View source

set_weights(
    weights
)

Sets the weights of the layer, from Numpy arrays.

__call__

View source

__call__(
    inputs, *args, **kwargs
)

Wraps call, applying pre- and post-processing steps.

Note:

• The following optional keyword arguments are reserved for specific uses:
  • training: Boolean scalar tensor of Python boolean indicating whether the call is meant for training or inference.
  • mask: Boolean input mask.
• If the layer's call method takes a mask argument (as some Keras layers do), its default value will be set to the mask generated for inputs by the previous layer (if input did come from a layer that generated a corresponding mask, i.e. if it came from a Keras layer with masking support.