def get_config(self):
config = {'filters': self.filters,
'kernel_size': self.kernel_size,
'strides': self.strides,
'padding': self.padding,
'data_format': self.data_format,
'dilation_rate': self.dilation_rate,
'activation': activations.serialize(self.activation),
'recurrent_activation': activations.serialize(
self.recurrent_activation),
'use_bias': self.use_bias,
'kernel_initializer': initializers.serialize(
self.kernel_initializer),
'recurrent_initializer': initializers.serialize(
self.recurrent_initializer),
'bias_initializer': initializers.serialize(self.bias_initializer),
'unit_forget_bias': self.unit_forget_bias,
'kernel_regularizer': regularizers.serialize(
self.kernel_regularizer),
'recurrent_regularizer': regularizers.serialize(
self.recurrent_regularizer),
'bias_regularizer': regularizers.serialize(self.bias_regularizer),
'kernel_constraint': constraints.serialize(
self.kernel_constraint),
'recurrent_constraint': constraints.serialize(
self.recurrent_constraint),
'bias_constraint': constraints.serialize(self.bias_constraint),
'dropout': self.dropout,
'recurrent_dropout': self.recurrent_dropout}
base_config = super(ConvLSTM2DCell, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'axis': self.axis,
'momentum': self.momentum,
'epsilon': self.epsilon,
'center': self.center,
'scale': self.scale,
'beta_initializer': initializers.serialize(self.beta_initializer),
'gamma_initializer': initializers.serialize(self.gamma_initializer),
'moving_mean_initializer':
initializers.serialize(self.moving_mean_initializer),
'moving_variance_initializer':
initializers.serialize(self.moving_variance_initializer),
'beta_regularizer': regularizers.serialize(self.beta_regularizer),
'gamma_regularizer': regularizers.serialize(self.gamma_regularizer),
'beta_constraint': constraints.serialize(self.beta_constraint),
'gamma_constraint': constraints.serialize(self.gamma_constraint)
}
# Only add TensorFlow-specific parameters if they are set, so as to preserve
# model compatibility with external Keras.
if self.renorm:
config['renorm'] = True
config['renorm_clipping'] = self.renorm_clipping
config['renorm_momentum'] = self.renorm_momentum
if self.virtual_batch_size is not None:
config['virtual_batch_size'] = self.virtual_batch_size
# Note: adjustment is not serializable.
if self.adjustment is not None:
logging.warning('The `adjustment` function of this `BatchNormalization` '
'layer cannot be serialized and has been omitted from '
'the layer config. It will not be included when '
're-creating the layer from the saved config.')
base_config = super(BatchNormalizationBase, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'filters':
self.filters,
'kernel_size':
self.kernel_size,
'strides':
self.strides,
'padding':
self.padding,
'data_format':
self.data_format,
'activation':
activations.serialize(self.activation),
'use_bias':
self.use_bias,
'kernel_initializer':
initializers.serialize(self.kernel_initializer),
'bias_initializer':
initializers.serialize(self.bias_initializer),
'kernel_regularizer':
regularizers.serialize(self.kernel_regularizer),
'bias_regularizer':
regularizers.serialize(self.bias_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'kernel_constraint':
constraints.serialize(self.kernel_constraint),
'bias_constraint':
constraints.serialize(self.bias_constraint),
'implementation':
self.implementation
}
base_config = super(LocallyConnected2D, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'axis': self.axis,
'epsilon': self.epsilon,
'center': self.center,
'scale': self.scale,
'beta_initializer': initializers.serialize(self.beta_initializer),
'gamma_initializer': initializers.serialize(self.gamma_initializer),
'beta_regularizer': regularizers.serialize(self.beta_regularizer),
'gamma_regularizer': regularizers.serialize(self.gamma_regularizer),
'beta_constraint': constraints.serialize(self.beta_constraint),
'gamma_constraint': constraints.serialize(self.gamma_constraint)
}
base_config = super(LayerNormalization, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'units': self.units,
'activation': activations.serialize(self.activation),
'use_bias': self.use_bias,
'kernel_initializer': initializers.serialize(self.kernel_initializer),
'bias_initializer': initializers.serialize(self.bias_initializer),
'kernel_regularizer': regularizers.serialize(self.kernel_regularizer),
'bias_regularizer': regularizers.serialize(self.bias_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'kernel_constraint': constraints.serialize(self.kernel_constraint),
'bias_constraint': constraints.serialize(self.bias_constraint)
}
base_config = super(Dense, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'alpha_initializer': initializers.serialize(self.alpha_initializer),
'alpha_regularizer': regularizers.serialize(self.alpha_regularizer),
'alpha_constraint': constraints.serialize(self.alpha_constraint),
'shared_axes': self.shared_axes
}
base_config = super(PReLU, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'units': self.units,
'kernel_initializer': initializers.serialize(self.kernel_initializer),
'recurrent_initializer':
initializers.serialize(self.recurrent_initializer),
'bias_initializer': initializers.serialize(self.bias_initializer),
'kernel_regularizer': regularizers.serialize(self.kernel_regularizer),
'recurrent_regularizer':
regularizers.serialize(self.recurrent_regularizer),
'bias_regularizer': regularizers.serialize(self.bias_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'kernel_constraint': constraints.serialize(self.kernel_constraint),
'recurrent_constraint':
constraints.serialize(self.recurrent_constraint),
'bias_constraint': constraints.serialize(self.bias_constraint)
}
base_config = super(CuDNNGRU, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
kernel_initializer = self.kernel_initializer
if isinstance(self.kernel_initializer, init_ops.Initializer):
kernel_initializer = initializers.serialize(self.kernel_initializer)
config = {
'output_dim': self.output_dim,
'kernel_initializer': kernel_initializer,
'scale': self.scale,
}
base_config = super(RandomFourierFeatures, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'input_dim': self.input_dim,
'output_dim': self.output_dim,
'embeddings_initializer':
initializers.serialize(self.embeddings_initializer),
'embeddings_regularizer':
regularizers.serialize(self.embeddings_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'embeddings_constraint':
constraints.serialize(self.embeddings_constraint),
'mask_zero': self.mask_zero,
'input_length': self.input_length
}
base_config = super(Embedding, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'units': self.units,
'activation': activations.serialize(self.activation),
# 'use_bias': self.use_bias,
'kernel_initializer':
initializers.serialize(self.kernel_initializer)
# 'bias_initializer': initializers.serialize(self.bias_initializer),
# 'kernel_regularizer': regularizers.serialize(self.kernel_regularizer),
# 'bias_regularizer': regularizers.serialize(self.bias_regularizer),
# 'activity_regularizer':
# regularizers.serialize(self.activity_regularizer),
# 'kernel_constraint': constraints.serialize(self.kernel_constraint),
# 'bias_constraint': constraints.serialize(self.bias_constraint)
}
base_config = super(RWNN, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
"num_units": self._num_units,
"use_peepholes": self._use_peepholes,
"cell_clip": self._cell_clip,
"initializer": initializers.serialize(self._initializer),
"num_proj": self._num_proj,
"proj_clip": self._proj_clip,
"num_unit_shards": self._num_unit_shards,
"num_proj_shards": self._num_proj_shards,
"forget_bias": self._forget_bias,
"state_is_tuple": self._state_is_tuple,
"activation": activations.serialize(self._activation),
"reuse": self._reuse,
}
base_config = super(TFLiteLSTMCell, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'input_dim': self.input_dim,
'output_dim': self.output_dim,
'embeddings_initializer':
initializers.serialize(self.embeddings_initializer),
'embeddings_regularizer':
regularizers.serialize(self.embeddings_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'embeddings_constraint':
constraints.serialize(self.embeddings_constraint),
'mask_zero': self.mask_zero,
'input_length': self.input_length
}
base_config = super(Embedding, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
"num_units": self._num_units,
"use_peepholes": self._use_peepholes,
"cell_clip": self._cell_clip,
"initializer": initializers.serialize(self._initializer),
"num_proj": self._num_proj,
"proj_clip": self._proj_clip,
"num_unit_shards": self._num_unit_shards,
"num_proj_shards": self._num_proj_shards,
"forget_bias": self._forget_bias,
"state_is_tuple": self._state_is_tuple,
"activation": activations.serialize(self._activation),
"reuse": self._reuse,
}
base_config = super(TFLiteLSTMCell, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
base_config = super(MultiHeadAttention, self).get_config()
base_config.update({
'output_dim':
self.output_dim,
'num_heads':
self.num_heads,
'negative_infinity':
self.negative_infinity,
'padding_value':
self.padding_value,
'kernel_initializer':
initializers.serialize(self.kernel_initializer),
'kernel_regularizer':
regularizers.serialize(self.kernel_regularizer),
'kernel_constraint':
constraints.serialize(self.kernel_constraint),
})
return base_config
def get_config(self):
# TODO this function is copied from Keras.Layers.Dense, and probably needs some further changes for TSRNNDense...?
config = {
'units':
self.units,
'activation':
activations.serialize(self.activation),
'use_bias':
self.use_bias,
'kernel_initializer':
initializers.serialize(self.kernel_initializer),
'kernel_regularizer':
regularizers.serialize(self.kernel_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'kernel_constraint':
constraints.serialize(self.kernel_constraint),
}
base_config = super(TSDense, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'tied_layer':
'',
'activation':
activations.serialize(self.activation),
'use_bias':
self.use_bias,
'bias_initializer':
initializers.serialize(self.bias_initializer),
'bias_regularizer':
regularizers.serialize(self.bias_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'bias_constraint':
constraints.serialize(self.bias_constraint),
'varName':
self.varName,
'varShape':
self.varShape
}
base_config = super(DenseTied, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def test_get_config(self, output_dim, initializer, scale, trainable):
rff_layer = kernel_layers.RandomFourierFeatures(
output_dim,
initializer,
scale=scale,
trainable=trainable,
name='random_fourier_features',
)
expected_initializer = initializer
if isinstance(initializer, init_ops.Initializer):
expected_initializer = initializers.serialize(initializer)
expected_config = {
'output_dim': output_dim,
'kernel_initializer': expected_initializer,
'scale': scale,
'name': 'random_fourier_features',
'trainable': trainable,
'dtype': None,
}
self.assertLen(expected_config, len(rff_layer.get_config()))
self.assertSameElements(
list(expected_config.items()), list(rff_layer.get_config().items()))
def get_config(self):
config = {
'activation':
activations.serialize(self.activation),
'use_bias':
self.use_bias,
'kernel_initializer':
self.kernel_initializer,
'bias_initializer':
initializers.serialize(self.bias_initializer),
'kernel_regularizer':
self.kernel_regularizer,
'bias_regularizer':
regularizers.serialize(self.bias_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'kernel_constraint':
self.kernel_constraint,
'bias_constraint':
constraints.serialize(self.bias_constraint)
}
base_config = super(_MultiTimeDelayLayer, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def test_get_config(self, output_dim, initializer, scale, trainable):
rff_layer = kernel_layers.RandomFourierFeatures(
output_dim,
initializer,
scale=scale,
trainable=trainable,
name='random_fourier_features',
)
expected_initializer = initializer
if isinstance(initializer, init_ops.Initializer):
expected_initializer = initializers.serialize(initializer)
expected_config = {
'output_dim': output_dim,
'kernel_initializer': expected_initializer,
'scale': scale,
'name': 'random_fourier_features',
'trainable': trainable,
'dtype': None,
}
self.assertLen(expected_config, len(rff_layer.get_config()))
self.assertSameElements(list(expected_config.items()),
list(rff_layer.get_config().items()))
def get_config(self):
config = {
'axis':
self.axis,
'epsilon':
self.epsilon,
'momentum':
self.momentum,
'center':
self.center,
'scale':
self.scale,
'beta_initializer':
initializers.serialize(self.beta_initializer),
'gamma_initializer':
initializers.serialize(self.gamma_initializer),
'mean_weights_initializer':
initializers.serialize(self.mean_weights_initializer),
'variance_weights_initializer':
initializers.serialize(self.variance_weights_initializer),
'moving_mean_initializer':
initializers.serialize(self.moving_mean_initializer),
'moving_variance_initializer':
initializers.serialize(self.moving_variance_initializer),
'beta_regularizer':
regularizers.serialize(self.beta_regularizer),
'gamma_regularizer':
regularizers.serialize(self.gamma_regularizer),
'mean_weights_regularizer':
regularizers.serialize(self.mean_weights_regularizer),
'variance_weights_regularizer':
regularizers.serialize(self.variance_weights_regularizer),
'beta_constraint':
constraints.serialize(self.beta_constraint),
'gamma_constraint':
constraints.serialize(self.gamma_constraint),
'mean_weights_constraints':
constraints.serialize(self.mean_weights_constraints),
'variance_weights_constraints':
constraints.serialize(self.variance_weights_constraints),
}
base_config = super(SwitchNormalization, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'initializer': initializers.serialize(self.initializer),
}
base_config = super(Gate, self).get_config()
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'units':
self.units,
'prior_memory_slots':
self.prior_memory_slots,
'posterior_memory_slots':
self.posterior_memory_slots,
'num_attention_heads':
self.num_attention_heads,
'activation':
wrap_activations_serialize(self.activation),
'recurrent_activation':
wrap_activations_serialize(self.recurrent_activation),
'mlp_activation':
wrap_activations_serialize(self.mlp_activation),
'forget_bias':
self.forget_bias,
'input_bias':
self.input_bias,
'sigma_bias':
self.sigma_bias,
'kernel_initializer':
initializers.serialize(self.kernel_initializer),
'recurrent_initializer':
initializers.serialize(self.recurrent_initializer),
'attention_initializer':
initializers.serialize(self.attention_initializer),
'mlp_initializer':
initializers.serialize(self.mlp_initializer),
'bias_initializer':
initializers.serialize(self.bias_initializer),
'kernel_regularizer':
regularizers.serialize(self.kernel_regularizer),
'recurrent_regularizer':
regularizers.serialize(self.recurrent_regularizer),
'attention_regularizer':
regularizers.serialize(self.attention_regularizer),
'mlp_regularizer':
regularizers.serialize(self.mlp_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'bias_regularizer':
regularizers.serialize(self.bias_regularizer),
'kernel_constraint':
constraints.serialize(self.kernel_constraint),
'recurrent_constraint':
constraints.serialize(self.recurrent_constraint),
'attention_constraint':
constraints.serialize(self.attention_constraint),
'mlp_constraint':
constraints.serialize(self.mlp_constraint),
'bias_constraint':
constraints.serialize(self.bias_constraint),
'dropout':
self.dropout,
'recurrent_dropout':
self.recurrent_dropout
}
base_config = super(TmpHierRNN, self).get_config()
del base_config['cell']
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'sub_units':
self.sub_units,
'sub_lstms':
self.sub_lstms,
'sub_activation':
activations.serialize(self.sub_activation),
'cake_activation':
activations.serialize(self.cake_activation),
'sub_use_bias':
self.sub_use_bias,
'cake_use_bias':
self.cake_use_bias,
'sub_kernel_initializer':
initializers.serialize(self.sub_kernel_initializer),
'cake_kernel_initializer':
initializers.serialize(self.cake_kernel_initializer),
'sub_recurrent_initializer':
initializers.serialize(self.sub_recurrent_initializer),
'cake_recurrent_initializer':
initializers.serialize(self.cake_recurrent_initializer),
'sub_bias_initializer':
initializers.serialize(self.sub_bias_initializer),
'cake_bias_initializer':
initializers.serialize(self.cake_bias_initializer),
'sub_unit_forget_bias':
self.sub_unit_forget_bias,
'cake_unit_forget_bias':
self.cake_unit_forget_bias,
'sub_kernel_regularizer':
regularizers.serialize(self.sub_kernel_regularizer),
'cake_kernel_regularizer':
regularizers.serialize(self.cake_kernel_regularizer),
'sub_recurrent_regularizer':
regularizers.serialize(self.sub_recurrent_regularizer),
'cake_recurrent_regularizer':
regularizers.serialize(self.cake_recurrent_regularizer),
'sub_bias_regularizer':
regularizers.serialize(self.sub_bias_regularizer),
'cake_bias_regularizer':
regularizers.serialize(self.cake_bias_regularizer),
'activity_regularizer':
regularizers.serialize(self.activity_regularizer),
'sub_kernel_constraint':
constraints.serialize(self.sub_kernel_constraint),
'cake_kernel_constraint':
constraints.serialize(self.cake_kernel_constraint),
'sub_recurrent_constraint':
constraints.serialize(self.sub_recurrent_constraint),
'cake_recurrent_constraint':
constraints.serialize(self.cake_recurrent_constraint),
'sub_bias_constraint':
constraints.serialize(self.sub_bias_constraint),
'cake_bias_constraint':
constraints.serialize(self.cake_bias_constraint),
'sub_dropout':
self.sub_dropout,
'cake_dropout':
self.cake_dropout,
'sub_recurrent_dropout':
self.sub_recurrent_dropout,
'cake_recurrent_dropout':
self.cake_recurrent_dropout,
'implementation':
self.implementation
}
base_config = super(JujubeCake, self).get_config()
del base_config['cell']
return dict(list(base_config.items()) + list(config.items()))
def get_config(self):
config = {
'units':
self.units,
'projection_units':
self.projection_units,
'use_feedback':
self.use_feedback,
'use_recurrent':
self.use_recurrent,
'activation':
activations.serialize(self.activation),
'projection_activation':
activations.serialize(self.projection_activation),
'use_bias':
self.use_bias,
'use_projection_bias':
self.use_projection_bias,
'kernel_initializer':
initializers.serialize(self.kernel_initializer),
'projection_initializer':
initializers.serialize(self.projection_initializer),
'recurrent_initializer':
initializers.serialize(self.recurrent_initializer),
'recurrent_initializer':
initializers.serialize(self.feedback_initializer),
'bias_initializer':
initializers.serialize(self.bias_initializer),
'bias_initializer':
initializers.serialize(self.projection_bias_initializer),
'kernel_regularizer':
regularizers.serialize(self.kernel_regularizer),
'projection_regularizer':
regularizers.serialize(self.projection_regularizer),
'recurrent_regularizer':
regularizers.serialize(self.recurrent_regularizer),
'feedback_regularizer':
regularizers.serialize(self.feedback_regularizer),
'bias_regularizer':
regularizers.serialize(self.bias_regularizer),
'projection_bias_regularizer':
regularizers.serialize(self.projection_bias_regularizer),
'kernel_constraint':
constraints.serialize(self.kernel_constraint),
'projection_constraint':
constraints.serialize(self.projection_constraint),
'recurrent_constraint':
constraints.serialize(self.recurrent_constraint),
'feedback_constraint':
constraints.serialize(self.feedback_constraint),
'bias_constraint':
constraints.serialize(self.bias_constraint),
'projection_bias_constraint':
constraints.serialize(self.projection_bias_constraint),
'dropout':
self.dropout,
'recurrent_dropout':
self.recurrent_dropout
}
base_config = super(Cell, self).get_config()
return dict(list(base_config.items()) + list(config.items()))