def add_variable(self,
name,
shape,
dtype=None,
trainable=True,
initializer=None,
regularizer=None):
if dtype is None:
dtype = self.dtype
existing_variables = set(tf_variables.global_variables())
with vs.variable_scope(self._scope, reuse=self.built
or self._reuse) as scope:
with ops.name_scope(scope.original_name_scope):
full_name = get_tensor_scope() + name
variable = vs.get_variable(name,
shape=shape,
initializer=initializer,
dtype=dtypes.as_dtype(dtype),
trainable=trainable
and self.trainable)
if variable in existing_variables:
# Work only if the layer is built
return variable
if regularizer:
raise NotImplementedError()
if trainable:
self._trainable_weights.append(variable)
else:
self._non_trainable_weights.append(variable)
return variable
def add_variable(self, name, shape, dtype=None, trainable=True,
initializer=None, regularizer=None):
if dtype is None:
dtype = self.dtype
existing_variables = set(tf_variables.global_variables())
with vs.variable_scope(self._scope,
reuse=self.built or self._reuse) as scope:
with ops.name_scope(scope.original_name_scope):
full_name = get_tensor_scope() + name
variable = vs.get_variable(name,
shape=shape,
initializer=initializer,
dtype=dtypes.as_dtype(dtype),
trainable=trainable and self.trainable)
if variable in existing_variables:
# Work only if the layer is built
return variable
if regularizer:
raise NotImplementedError()
if trainable:
self._trainable_weights.append(variable)
else:
self._non_trainable_weights.append(variable)
return variable
def __call__(self, inputs, *args, **kwargs):
with vs.variable_scope(self._scope, reuse=self.built
or self._reuse) as scope:
with ops.name_scope(scope.original_name_scope):
if not self.built:
input_shapes = [
x.get_shape() for x in nest.flatten(inputs)
]
if len(input_shapes) == 1:
self.build(input_shapes[0])
else:
self.build(input_shapes)
outputs = self.call(inputs, *args, **kwargs)
# # Apply activity regularization.
# # Note that it should be applied every time the layer creates a new
# # output, since it is output-specific.
# if hasattr(self, 'activity_regularizer') and self.activity_regularizer:
# output_list = _to_list(outputs)
# for output in output_list:
# with ops.name_scope('ActivityRegularizer'):
# activity_regularization = self.activity_regularizer(output)
# self.add_loss(activity_regularization)
# _add_elements_to_collection(
# activity_regularization, ops.GraphKeys.REGULARIZATION_LOSSES)
# Update global default collections.
_add_elements_to_collection(self.updates, ops.GraphKeys.UPDATE_OPS)
self.built = True
return outputs
def __call__(self, inputs, *args, **kwargs):
with vs.variable_scope(self._scope,
reuse=self.built or self._reuse) as scope:
with ops.name_scope(scope.original_name_scope):
if not self.built:
input_shapes = [x.get_shape()
for x in nest.flatten(inputs)]
if len(input_shapes) == 1:
self.build(input_shapes[0])
else:
self.build(input_shapes)
outputs = self.call(inputs, *args, **kwargs)
# # Apply activity regularization.
# # Note that it should be applied every time the layer creates a new
# # output, since it is output-specific.
# if hasattr(self, 'activity_regularizer') and self.activity_regularizer:
# output_list = _to_list(outputs)
# for output in output_list:
# with ops.name_scope('ActivityRegularizer'):
# activity_regularization = self.activity_regularizer(output)
# self.add_loss(activity_regularization)
# _add_elements_to_collection(
# activity_regularization, ops.GraphKeys.REGULARIZATION_LOSSES)
# Update global default collections.
_add_elements_to_collection(self.updates, ops.GraphKeys.UPDATE_OPS)
self.built = True
return outputs