def _test_body():
# pylint: disable=protected-access
if context.executing_eagerly():
beam_search_decoder._check_batch_beam(t, batch_size, beam_width)
else:
with self.cached_session():
check_op = beam_search_decoder._check_batch_beam(
t, batch_size, beam_width)
self.evaluate(check_op)
def _test_body():
# pylint: disable=protected-access
if context.executing_eagerly():
beam_search_decoder._check_batch_beam(t, batch_size,
beam_width)
else:
with self.cached_session():
check_op = beam_search_decoder._check_batch_beam(
t, batch_size, beam_width)
self.evaluate(check_op)
def _test_array_shape_dynamic_checks(self, static_shape, dynamic_shape,
batch_size, beam_width, is_valid=True):
t = array_ops.placeholder_with_default(
np.random.randn(*static_shape).astype(np.float32),
shape=dynamic_shape)
batch_size = array_ops.constant(batch_size)
check_op = beam_search_decoder._check_batch_beam(t, batch_size, beam_width) # pylint: disable=protected-access
with self.cached_session() as sess:
if is_valid:
sess.run(check_op)
else:
with self.assertRaises(errors.InvalidArgumentError):
sess.run(check_op)
def _test_array_shape_dynamic_checks(self, static_shape, dynamic_shape,
batch_size, beam_width, is_valid=True):
t = array_ops.placeholder_with_default(
np.random.randn(*static_shape).astype(np.float32),
shape=dynamic_shape)
batch_size = array_ops.constant(batch_size)
check_op = beam_search_decoder._check_batch_beam(t, batch_size, beam_width) # pylint: disable=protected-access
with self.cached_session() as sess:
if is_valid:
sess.run(check_op)
else:
with self.assertRaises(errors.InvalidArgumentError):
sess.run(check_op)
def _maybe_sort_array_beams(self, t, parent_ids, sequence_length):
"""Maybe sorts beams within a `TensorArray`.
Args:
t: A `TensorArray` of size `max_time` that contains `Tensor`s of shape
`[batch_size, beam_width, s]` or `[batch_size * beam_width, s]` where
`s` is the depth shape.
parent_ids: The parent ids of shape `[max_time, batch_size, beam_width]`.
sequence_length: The sequence length of shape `[batch_size, beam_width]`.
Returns:
A `TensorArray` where beams are sorted in each `Tensor` or `t` itself if
it is not a `TensorArray` or does not meet shape requirements.
"""
if not isinstance(t, tf.TensorArray):
return t
# pylint: disable=protected-access
if (not t._infer_shape or not t._element_shape
or t._element_shape[0].ndims is None
or t._element_shape[0].ndims < 1):
shape = (t._element_shape[0] if t._infer_shape and t._element_shape
else tf.TensorShape(None))
tf.logger.warn(
"The TensorArray %s in the cell state is not amenable to "
"sorting based on the beam search result. For a "
"TensorArray to be sorted, its elements shape must be "
"defined and have at least a rank of 1, but saw shape: %s" %
(t.handle.name, shape))
return t
# shape = t._element_shape[0]
# pylint: enable=protected-access
# if not _check_static_batch_beam_maybe(
# shape, tensor_util.constant_value(self._batch_size), self._beam_width):
# return t
t = t.stack()
with tf.control_dependencies(
[_check_batch_beam(t, self._batch_size, self._beam_width)]):
return gather_tree_from_array(t, parent_ids, sequence_length)