def __call__(self, inputs, initial_state=None, initial_readout=None, ground_truth=None, **kwargs):
req_num_inputs = 1 + self.num_states
inputs = _to_list(inputs)
inputs = inputs[:]
if len(inputs) == 1:
if initial_state is not None:
if type(initial_state) is list:
inputs += initial_state
else:
inputs.append(initial_state)
else:
if self.readout:
initial_state = self._get_optional_input_placeholder('initial_state', self.num_states - 1)
else:
initial_state = self._get_optional_input_placeholder('initial_state', self.num_states)
inputs += _to_list(initial_state)
if self.readout:
if initial_readout is None:
initial_readout = self._get_optional_input_placeholder('initial_readout')
inputs.append(initial_readout)
if self.teacher_force:
req_num_inputs += 1
if ground_truth is None:
ground_truth = self._get_optional_input_placeholder('ground_truth')
inputs.append(ground_truth)
assert len(inputs) == req_num_inputs, "Required " + str(req_num_inputs) + " inputs, received " + str(len(inputs)) + "."
with K.name_scope(self.name):
if not self.built:
self.build(K.int_shape(inputs[0]))
if self._initial_weights is not None:
self.set_weights(self._initial_weights)
del self._initial_weights
self._initial_weights = None
previous_mask = _collect_previous_mask(inputs[:1])
user_kwargs = kwargs.copy()
if not _is_all_none(previous_mask):
if 'mask' in inspect.getargspec(self.call).args:
if 'mask' not in kwargs:
kwargs['mask'] = previous_mask
input_shape = _collect_input_shape(inputs)
output = self.call(inputs, **kwargs)
output_mask = self.compute_mask(inputs[0], previous_mask)
output_shape = self.compute_output_shape(input_shape[0])
self._add_inbound_node(input_tensors=inputs, output_tensors=output,
input_masks=previous_mask, output_masks=output_mask,
input_shapes=input_shape, output_shapes=output_shape,
arguments=user_kwargs)
if hasattr(self, 'activity_regularizer') and self.activity_regularizer is not None:
regularization_losses = [self.activity_regularizer(x) for x in _to_list(output)]
self.add_loss(regularization_losses, _to_list(inputs))
return output
def call(self, inputs, states, constants, training=None):
"""Complete attentive cell transformation.
"""
attended = to_list(constants, allow_tuple=True)
# NOTE: `K.rnn` will pass constants as a tuple and `_collect_previous_mask`
# returns `None` if passed a tuple of tensors, hence `to_list` above!
# We also make `attended` and `attended_mask` always lists for uniformity:
attended_mask = to_list(_collect_previous_mask(attended))
cell_states = states[:self._num_wrapped_states]
attention_states = states[self._num_wrapped_states:]
if self.attend_after:
call = self._call_attend_after
else:
call = self._call_attend_before
return call(inputs=inputs,
cell_states=cell_states,
attended=attended,
attention_states=attention_states,
attended_mask=attended_mask,
training=training)
def call(self, inputs, states, constants, training=None):
"""Complete attentive cell transformation.
"""
attended = constants
attended_mask = _collect_previous_mask(attended)
# attended and mask are always lists for uniformity:
if not isinstance(attended_mask, list):
attended_mask = [attended_mask]
cell_states = states[:self._num_wrapped_states]
attention_states = states[self._num_wrapped_states:]
if self.attend_after:
call = self._call_attend_after
else:
call = self._call_attend_before
return call(inputs=inputs,
cell_states=cell_states,
attended=attended,
attention_states=attention_states,
attended_mask=attended_mask,
training=training)
def __call__(self, inputs, **kwargs):
"""
Overriding the Layer's __call__ method for graph wrappers.
The overriding is mostly based on the __call__ code from
'keras/engine/base_layer.py' for the moment, with some changes
to handle graphs.
"""
# If arguments are 'keras-style' arguments, let keras to the job
if K.is_tensor(inputs) or (isinstance(inputs, list)
and not isinstance(inputs, GraphWrapper)):
output = super(GraphLayer, self).__call__(inputs, **kwargs)
else:
if isinstance(inputs,
list) and not isinstance(inputs, GraphWrapper):
inputs = inputs[:]
with K.name_scope(self.name):
# Build the layer
if not self.built:
# Check the comatibilty of inputs for each inputs (some can
# be GraphWrapper)
if isinstance(inputs, list) and not isinstance(
inputs, GraphWrapper):
for i in inputs:
self.assert_input_compatibility(i)
# Collect input shapes to build layer.
input_shapes = []
if isinstance(inputs, GraphWrapper):
inputs = [inputs]
for x_elem in inputs:
if hasattr(x_elem, '_keras_shape'):
# For a GraphWrapper, _keras_shape is a GraphShape
# object
input_shapes.append(x_elem._keras_shape)
elif hasattr(K, 'int_shape'):
input_shapes.append(K.int_shape(x_elem))
else:
raise ValueError(
'You tried to call layer "' + self.name +
'". This layer has no information'
' about its expected input shape, '
'and thus cannot be built. '
'You can build it manually via: '
'`layer.build(batch_input_shape)`')
self.build(unpack_singleton(input_shapes))
self._built = True
# Load weights that were specified at layer instantiation.
if self._initial_weights is not None:
self.set_weights(self._initial_weights)
# Raise exceptions in case the input is not compatible
# with the input_spec set at build time.
if isinstance(inputs,
list) and not isinstance(inputs, GraphWrapper):
for i in inputs:
self.assert_input_compatibility(i)
# Handle mask propagation.
previous_mask = _collect_previous_mask(inputs)
user_kwargs = copy.copy(kwargs)
if not is_all_none(previous_mask):
# The previous layer generated a mask.
if has_arg(self.call, 'mask'):
if 'mask' not in kwargs:
# If mask is explicitly passed to __call__,
# we should override the default mask.
kwargs['mask'] = previous_mask
# Handle automatic shape inference (only useful for Theano).
input_shape = _collect_input_shape(inputs)
# Actually call the layer,
# collecting output(s), mask(s), and shape(s).
# Note that inpputs can hold graph wrappers now
output = self.call(unpack_singleton(inputs), **kwargs)
output_mask = self.compute_mask(inputs, previous_mask)
# If the layer returns tensors from its inputs, unmodified,
# we copy them to avoid loss of tensor metadata.
# output_ls = to_list(output)
if isinstance(output, GraphWrapper):
output_ls = [output]
# Unpack wrappers
inputs_ls = list(
chain.from_iterable(
[i for i in inputs if isinstance(i, GraphWrapper)]))
inputs_ls += [
i for i in inputs if not isinstance(i, GraphWrapper)
]
# Unpack adjacency and nodes
inpadj_ls = list(
chain.from_iterable([
to_list(i.adjacency) for i in inputs
if isinstance(i, GraphWrapper)
]))
inpnod_ls = to_list(
[i.nodes for i in inputs if isinstance(i, GraphWrapper)])
output_ls_copy = []
adj_ls = []
for x in output_ls:
if K.is_tensor(x) and x in inputs_ls:
x = K.identity(x)
# Apply adjacency-wise and node-wise identity
elif isinstance(x, GraphWrapper):
# Store changed or copy of unchanged adjacency matrices
adj_ls.clear()
for adj in to_list(x.adjacency):
if adj in inpadj_ls:
adj_ls.append(K.identity(adj))
else:
adj_ls.append(adj)
# Assign to output graph
x.adjacency = unpack_singleton(adj_ls)
# Store unchanged nodes
if x.nodes in inpnod_ls:
x.nodes = K.identity(x.nodes)
output_ls_copy.append(x)
output = unpack_singleton(output_ls_copy)
# Inferring the output shape is only relevant for Theano.
if all([s is not None for s in to_list(input_shape)]):
output_shape = self.compute_output_shape(input_shape)
else:
if isinstance(input_shape, list):
output_shape = [None for _ in input_shape]
else:
output_shape = None
if (not isinstance(output_mask, (list, tuple))
and len(output_ls) > 1):
# Augment the mask to match the length of the output.
output_mask = [output_mask] * len(output_ls)
# Add an inbound node to the layer, so that it keeps track
# of the call and of all new variables created during the call.
# This also updates the layer history of the output tensor(s).
# If the input tensor(s) had not previous Keras history,
# this does nothing.
self._add_inbound_node(input_tensors=inputs_ls,
output_tensors=unpack_singleton(output),
input_masks=previous_mask,
output_masks=output_mask,
input_shapes=input_shape,
output_shapes=output_shape,
arguments=user_kwargs)
# Apply activity regularizer if any:
if (hasattr(self, 'activity_regularizer')
and self.activity_regularizer is not None):
with K.name_scope('activity_regularizer'):
regularization_losses = [
self.activity_regularizer(x)
for x in to_list(output)
]
self.add_loss(regularization_losses,
inputs=to_list(inputs))
return output