def graph_fn():
label_map_proto_file = os.path.join(self.get_temp_dir(), 'labelmap.pbtxt')
self._create_label_map(label_map_proto_file)
decoder = tf_sequence_example_decoder.TfSequenceExampleDecoder(
label_map_proto_file=label_map_proto_file)
sequence_example_serialized = seq_example_util.make_sequence_example(
dataset_name='video_dataset',
video_id='video',
encoded_images=encoded_images,
image_height=image_height,
image_width=image_width,
image_format='JPEG',
image_source_ids=[str(i) for i in range(num_frames)],
is_annotated=[[1], [1], [1], [1]],
bboxes=[
[[0., 0., 1., 1.]], # Frame 0.
[[0.2, 0.2, 1., 1.],
[0., 0., 1., 1.]], # Frame 1.
[[0., 0., 1., 1.], # Frame 2.
[0.1, 0.1, 0.2, 0.2]],
[[]], # Frame 3.
],
label_strings=[
['fox'], # Frame 0. Fox will be filtered out.
['fox', 'dog'], # Frame 1. Fox will be filtered out.
['dog', 'cat'], # Frame 2.
[], # Frame 3
]).SerializeToString()
example_string_tensor = tf.convert_to_tensor(sequence_example_serialized)
return decoder.decode(example_string_tensor)
def graph_fn():
sequence_example_serialized = seq_example_util.make_sequence_example(
dataset_name='video_dataset',
video_id='video',
encoded_images=encoded_images,
image_height=image_height,
image_width=image_width,
image_format='JPEG',
image_source_ids=[str(i) for i in range(num_frames)],
bboxes=[
[[]],
[[]],
[[]],
[[]]
],
label_strings=[
[],
[],
[],
[]
]).SerializeToString()
example_string_tensor = tf.convert_to_tensor(sequence_example_serialized)
label_map_proto_file = os.path.join(self.get_temp_dir(), 'labelmap.pbtxt')
self._create_label_map(label_map_proto_file)
decoder = tf_sequence_example_decoder.TfSequenceExampleDecoder(
label_map_proto_file=label_map_proto_file)
return decoder.decode(example_string_tensor)
def build(input_reader_config):
"""Builds a tensor dictionary based on the InputReader config.
Args:
input_reader_config: A input_reader_pb2.InputReader object.
Returns:
A tensor dict based on the input_reader_config.
Raises:
ValueError: On invalid input reader proto.
ValueError: If no input paths are specified.
"""
if not isinstance(input_reader_config, input_reader_pb2.InputReader):
raise ValueError('input_reader_config not of type '
'input_reader_pb2.InputReader.')
if input_reader_config.WhichOneof(
'input_reader') == 'tf_record_input_reader':
config = input_reader_config.tf_record_input_reader
if not config.input_path:
raise ValueError('At least one input path must be specified in '
'`input_reader_config`.')
_, string_tensor = parallel_reader.parallel_read(
config.input_path[:], # Convert `RepeatedScalarContainer` to list.
reader_class=tf.TFRecordReader,
num_epochs=(input_reader_config.num_epochs
if input_reader_config.num_epochs else None),
num_readers=input_reader_config.num_readers,
shuffle=input_reader_config.shuffle,
dtypes=[tf.string, tf.string],
capacity=input_reader_config.queue_capacity,
min_after_dequeue=input_reader_config.min_after_dequeue)
label_map_proto_file = None
if input_reader_config.HasField('label_map_path'):
label_map_proto_file = input_reader_config.label_map_path
input_type = input_reader_config.input_type
if input_type == input_reader_pb2.InputType.Value('TF_EXAMPLE'):
decoder = tf_example_decoder.TfExampleDecoder(
load_instance_masks=input_reader_config.load_instance_masks,
instance_mask_type=input_reader_config.mask_type,
label_map_proto_file=label_map_proto_file,
load_context_features=input_reader_config.load_context_features
)
return decoder.decode(string_tensor)
elif input_type == input_reader_pb2.InputType.Value(
'TF_SEQUENCE_EXAMPLE'):
decoder = tf_sequence_example_decoder.TfSequenceExampleDecoder(
label_map_proto_file=label_map_proto_file,
load_context_features=input_reader_config.load_context_features
)
return decoder.decode(string_tensor)
raise ValueError('Unsupported input_type.')
raise ValueError('Unsupported input_reader_config.')
def build(input_reader_config):
"""Builds a DataDecoder based only on the open source config proto.
Args:
input_reader_config: An input_reader_pb2.InputReader object.
Returns:
A DataDecoder based on the input_reader_config.
Raises:
ValueError: On invalid input reader proto.
"""
if not isinstance(input_reader_config, input_reader_pb2.InputReader):
raise ValueError('input_reader_config not of type '
'input_reader_pb2.InputReader.')
if input_reader_config.WhichOneof(
'input_reader') == 'tf_record_input_reader':
label_map_proto_file = None
if input_reader_config.HasField('label_map_path'):
label_map_proto_file = input_reader_config.label_map_path
input_type = input_reader_config.input_type
if input_type == input_reader_pb2.InputType.Value('TF_EXAMPLE'):
decoder = tf_example_decoder.TfExampleDecoder(
load_instance_masks=input_reader_config.load_instance_masks,
load_multiclass_scores=input_reader_config.
load_multiclass_scores,
load_context_features=input_reader_config.
load_context_features,
instance_mask_type=input_reader_config.mask_type,
label_map_proto_file=label_map_proto_file,
use_display_name=input_reader_config.use_display_name,
num_additional_channels=input_reader_config.
num_additional_channels,
num_keypoints=input_reader_config.num_keypoints,
expand_hierarchy_labels=input_reader_config.
expand_labels_hierarchy,
load_dense_pose=input_reader_config.load_dense_pose,
load_track_id=input_reader_config.load_track_id)
return decoder
elif input_type == input_reader_pb2.InputType.Value(
'TF_SEQUENCE_EXAMPLE'):
decoder = tf_sequence_example_decoder.TfSequenceExampleDecoder(
label_map_proto_file=label_map_proto_file,
load_context_features=input_reader_config.
load_context_features,
load_context_image_ids=input_reader_config.
load_context_image_ids)
return decoder
raise ValueError('Unsupported input_type in config.')
raise ValueError('Unsupported input_reader_config.')