def test_raises_error_with_no_input_paths(self):
input_reader_text_proto = """
shuffle: false
num_readers: 1
load_instance_masks: true
"""
input_reader_proto = input_reader_pb2.InputReader()
text_format.Merge(input_reader_text_proto, input_reader_proto)
with self.assertRaises(ValueError):
dataset_builder.build(input_reader_proto)
def test_build_tf_record_input_reader(self):
tf_record_path = self.create_tf_record()
input_reader_text_proto = """
shuffle: false
num_readers: 1
tf_record_input_reader {{
input_path: '{0}'
}}
""".format(tf_record_path)
input_reader_proto = input_reader_pb2.InputReader()
text_format.Merge(input_reader_text_proto, input_reader_proto)
tensor_dict = dataset_util.make_initializable_iterator(
dataset_builder.build(input_reader_proto)).get_next()
sv = tf.train.Supervisor(logdir=self.get_temp_dir())
with sv.prepare_or_wait_for_session() as sess:
sv.start_queue_runners(sess)
output_dict = sess.run(tensor_dict)
self.assertTrue(
fields.InputDataFields.groundtruth_instance_masks not in output_dict)
self.assertEquals((4, 5, 3),
output_dict[fields.InputDataFields.image].shape)
self.assertEquals([2],
output_dict[fields.InputDataFields.groundtruth_classes])
self.assertEquals(
(1, 4), output_dict[fields.InputDataFields.groundtruth_boxes].shape)
self.assertAllEqual(
[0.0, 0.0, 1.0, 1.0],
output_dict[fields.InputDataFields.groundtruth_boxes][0])
def test_build_tf_record_input_reader_with_batch_size_two_and_masks(self):
tf_record_path = self.create_tf_record()
input_reader_text_proto = """
shuffle: false
num_readers: 1
load_instance_masks: true
tf_record_input_reader {{
input_path: '{0}'
}}
""".format(tf_record_path)
input_reader_proto = input_reader_pb2.InputReader()
text_format.Merge(input_reader_text_proto, input_reader_proto)
def one_hot_class_encoding_fn(tensor_dict):
tensor_dict[fields.InputDataFields.groundtruth_classes] = tf.one_hot(
tensor_dict[fields.InputDataFields.groundtruth_classes] - 1, depth=3)
return tensor_dict
tensor_dict = dataset_builder.make_initializable_iterator(
dataset_builder.build(
input_reader_proto,
transform_input_data_fn=one_hot_class_encoding_fn,
batch_size=2)).get_next()
sv = tf.train.Supervisor(logdir=self.get_temp_dir())
with sv.prepare_or_wait_for_session() as sess:
sv.start_queue_runners(sess)
output_dict = sess.run(tensor_dict)
self.assertAllEqual(
[2, 1, 4, 5],
output_dict[fields.InputDataFields.groundtruth_instance_masks].shape)
def test_build_tf_record_input_reader_with_batch_size_two(self):
tf_record_path = self.create_tf_record()
input_reader_text_proto = """
shuffle: false
num_readers: 1
tf_record_input_reader {{
input_path: '{0}'
}}
""".format(tf_record_path)
input_reader_proto = input_reader_pb2.InputReader()
text_format.Merge(input_reader_text_proto, input_reader_proto)
def one_hot_class_encoding_fn(tensor_dict):
tensor_dict[fields.InputDataFields.groundtruth_classes] = tf.one_hot(
tensor_dict[fields.InputDataFields.groundtruth_classes] - 1, depth=3)
return tensor_dict
tensor_dict = dataset_util.make_initializable_iterator(
dataset_builder.build(
input_reader_proto,
transform_input_data_fn=one_hot_class_encoding_fn,
batch_size=2,
max_num_boxes=2,
num_classes=3,
spatial_image_shape=[4, 5])).get_next()
sv = tf.train.Supervisor(logdir=self.get_temp_dir())
with sv.prepare_or_wait_for_session() as sess:
sv.start_queue_runners(sess)
output_dict = sess.run(tensor_dict)
self.assertAllEqual([2, 4, 5, 3],
output_dict[fields.InputDataFields.image].shape)
self.assertAllEqual([2, 2, 3],
output_dict[fields.InputDataFields.groundtruth_classes].
shape)
self.assertAllEqual([2, 2, 4],
output_dict[fields.InputDataFields.groundtruth_boxes].
shape)
self.assertAllEqual(
[[[0.0, 0.0, 1.0, 1.0],
[0.0, 0.0, 0.0, 0.0]],
[[0.0, 0.0, 1.0, 1.0],
[0.0, 0.0, 0.0, 0.0]]],
output_dict[fields.InputDataFields.groundtruth_boxes])
def test_build_tf_record_input_reader_and_load_instance_masks(self):
tf_record_path = self.create_tf_record()
input_reader_text_proto = """
shuffle: false
num_readers: 1
load_instance_masks: true
tf_record_input_reader {{
input_path: '{0}'
}}
""".format(tf_record_path)
input_reader_proto = input_reader_pb2.InputReader()
text_format.Merge(input_reader_text_proto, input_reader_proto)
tensor_dict = dataset_builder.make_initializable_iterator(
dataset_builder.build(input_reader_proto, batch_size=1)).get_next()
sv = tf.train.Supervisor(logdir=self.get_temp_dir())
with sv.prepare_or_wait_for_session() as sess:
sv.start_queue_runners(sess)
output_dict = sess.run(tensor_dict)
self.assertAllEqual(
(1, 1, 4, 5),
output_dict[fields.InputDataFields.groundtruth_instance_masks].shape)
def test_build_tf_record_input_reader_with_additional_channels(self):
tf_record_path = self.create_tf_record(has_additional_channels=True)
input_reader_text_proto = """
shuffle: false
num_readers: 1
tf_record_input_reader {{
input_path: '{0}'
}}
""".format(tf_record_path)
input_reader_proto = input_reader_pb2.InputReader()
text_format.Merge(input_reader_text_proto, input_reader_proto)
tensor_dict = dataset_util.make_initializable_iterator(
dataset_builder.build(
input_reader_proto, batch_size=2,
num_additional_channels=2)).get_next()
sv = tf.train.Supervisor(logdir=self.get_temp_dir())
with sv.prepare_or_wait_for_session() as sess:
sv.start_queue_runners(sess)
output_dict = sess.run(tensor_dict)
self.assertEquals((2, 4, 5, 5),
output_dict[fields.InputDataFields.image].shape)
def _train_input_fn(params=None):
"""Returns `features` and `labels` tensor dictionaries for training.
Args:
params: Parameter dictionary passed from the estimator.
Returns:
features: Dictionary of feature tensors.
features[fields.InputDataFields.image] is a [batch_size, H, W, C]
float32 tensor with preprocessed images.
features[HASH_KEY] is a [batch_size] int32 tensor representing unique
identifiers for the images.
features[fields.InputDataFields.true_image_shape] is a [batch_size, 3]
int32 tensor representing the true image shapes, as preprocessed
images could be padded.
labels: Dictionary of groundtruth tensors.
labels[fields.InputDataFields.num_groundtruth_boxes] is a [batch_size]
int32 tensor indicating the number of groundtruth boxes.
labels[fields.InputDataFields.groundtruth_boxes] is a
[batch_size, num_boxes, 4] float32 tensor containing the corners of
the groundtruth boxes.
labels[fields.InputDataFields.groundtruth_classes] is a
[batch_size, num_boxes, num_classes] float32 one-hot tensor of
classes.
labels[fields.InputDataFields.groundtruth_weights] is a
[batch_size, num_boxes] float32 tensor containing groundtruth weights
for the boxes.
-- Optional --
labels[fields.InputDataFields.groundtruth_instance_masks] is a
[batch_size, num_boxes, H, W] float32 tensor containing only binary
values, which represent instance masks for objects.
labels[fields.InputDataFields.groundtruth_keypoints] is a
[batch_size, num_boxes, num_keypoints, 2] float32 tensor containing
keypoints for each box.
Raises:
TypeError: if the `train_config` or `train_input_config` are not of the
correct type.
"""
if not isinstance(train_config, train_pb2.TrainConfig):
raise TypeError('For training mode, the `train_config` must be a '
'train_pb2.TrainConfig.')
if not isinstance(train_input_config, input_reader_pb2.InputReader):
raise TypeError('The `train_input_config` must be a '
'input_reader_pb2.InputReader.')
if not isinstance(model_config, model_pb2.DetectionModel):
raise TypeError('The `model_config` must be a '
'model_pb2.DetectionModel.')
data_augmentation_options = [
preprocessor_builder.build(step)
for step in train_config.data_augmentation_options
]
data_augmentation_fn = functools.partial(
augment_input_data, data_augmentation_options=data_augmentation_options)
model = model_builder.build(model_config, is_training=True)
image_resizer_config = config_util.get_image_resizer_config(model_config)
image_resizer_fn = image_resizer_builder.build(image_resizer_config)
transform_data_fn = functools.partial(
transform_input_data, model_preprocess_fn=model.preprocess,
image_resizer_fn=image_resizer_fn,
num_classes=config_util.get_number_of_classes(model_config),
data_augmentation_fn=data_augmentation_fn)
dataset = dataset_builder.build(
train_input_config,
transform_input_data_fn=transform_data_fn,
batch_size=params['batch_size'] if params else train_config.batch_size,
max_num_boxes=train_config.max_number_of_boxes,
num_classes=config_util.get_number_of_classes(model_config),
spatial_image_shape=config_util.get_spatial_image_size(
image_resizer_config))
tensor_dict = dataset_util.make_initializable_iterator(dataset).get_next()
hash_from_source_id = tf.string_to_hash_bucket_fast(
tensor_dict[fields.InputDataFields.source_id], HASH_BINS)
features = {
fields.InputDataFields.image: tensor_dict[fields.InputDataFields.image],
HASH_KEY: tf.cast(hash_from_source_id, tf.int32),
fields.InputDataFields.true_image_shape: tensor_dict[
fields.InputDataFields.true_image_shape]
}
labels = {
fields.InputDataFields.num_groundtruth_boxes: tensor_dict[
fields.InputDataFields.num_groundtruth_boxes],
fields.InputDataFields.groundtruth_boxes: tensor_dict[
fields.InputDataFields.groundtruth_boxes],
fields.InputDataFields.groundtruth_classes: tensor_dict[
fields.InputDataFields.groundtruth_classes],
fields.InputDataFields.groundtruth_weights: tensor_dict[
fields.InputDataFields.groundtruth_weights]
}
if fields.InputDataFields.groundtruth_keypoints in tensor_dict:
labels[fields.InputDataFields.groundtruth_keypoints] = tensor_dict[
fields.InputDataFields.groundtruth_keypoints]
if fields.InputDataFields.groundtruth_instance_masks in tensor_dict:
labels[fields.InputDataFields.groundtruth_instance_masks] = tensor_dict[
fields.InputDataFields.groundtruth_instance_masks]
return features, labels
def _eval_input_fn(params=None):
"""Returns `features` and `labels` tensor dictionaries for evaluation.
Args:
params: Parameter dictionary passed from the estimator.
Returns:
features: Dictionary of feature tensors.
features[fields.InputDataFields.image] is a [1, H, W, C] float32 tensor
with preprocessed images.
features[HASH_KEY] is a [1] int32 tensor representing unique
identifiers for the images.
features[fields.InputDataFields.true_image_shape] is a [1, 3]
int32 tensor representing the true image shapes, as preprocessed
images could be padded.
features[fields.InputDataFields.original_image] is a [1, H', W', C]
float32 tensor with the original image.
labels: Dictionary of groundtruth tensors.
labels[fields.InputDataFields.groundtruth_boxes] is a [1, num_boxes, 4]
float32 tensor containing the corners of the groundtruth boxes.
labels[fields.InputDataFields.groundtruth_classes] is a
[num_boxes, num_classes] float32 one-hot tensor of classes.
labels[fields.InputDataFields.groundtruth_area] is a [1, num_boxes]
float32 tensor containing object areas.
labels[fields.InputDataFields.groundtruth_is_crowd] is a [1, num_boxes]
bool tensor indicating if the boxes enclose a crowd.
labels[fields.InputDataFields.groundtruth_difficult] is a [1, num_boxes]
int32 tensor indicating if the boxes represent difficult instances.
-- Optional --
labels[fields.InputDataFields.groundtruth_instance_masks] is a
[1, num_boxes, H, W] float32 tensor containing only binary values,
which represent instance masks for objects.
Raises:
TypeError: if the `eval_config` or `eval_input_config` are not of the
correct type.
"""
del params
if not isinstance(eval_config, eval_pb2.EvalConfig):
raise TypeError('For eval mode, the `eval_config` must be a '
'train_pb2.EvalConfig.')
if not isinstance(eval_input_config, input_reader_pb2.InputReader):
raise TypeError('The `eval_input_config` must be a '
'input_reader_pb2.InputReader.')
if not isinstance(model_config, model_pb2.DetectionModel):
raise TypeError('The `model_config` must be a '
'model_pb2.DetectionModel.')
num_classes = config_util.get_number_of_classes(model_config)
model = model_builder.build(model_config, is_training=False)
image_resizer_config = config_util.get_image_resizer_config(
model_config)
image_resizer_fn = image_resizer_builder.build(image_resizer_config)
transform_data_fn = functools.partial(
transform_input_data,
model_preprocess_fn=model.preprocess,
image_resizer_fn=image_resizer_fn,
num_classes=num_classes,
data_augmentation_fn=None,
retain_original_image=True)
dataset = dataset_builder.build(
eval_input_config, transform_input_data_fn=transform_data_fn)
input_dict = dataset_util.make_initializable_iterator(
dataset).get_next()
hash_from_source_id = tf.string_to_hash_bucket_fast(
input_dict[fields.InputDataFields.source_id], HASH_BINS)
features = {
fields.InputDataFields.image:
input_dict[fields.InputDataFields.image],
fields.InputDataFields.original_image:
input_dict[fields.InputDataFields.original_image],
HASH_KEY:
tf.cast(hash_from_source_id, tf.int32),
fields.InputDataFields.true_image_shape:
input_dict[fields.InputDataFields.true_image_shape]
}
labels = {
fields.InputDataFields.groundtruth_boxes:
input_dict[fields.InputDataFields.groundtruth_boxes],
fields.InputDataFields.groundtruth_classes:
input_dict[fields.InputDataFields.groundtruth_classes],
fields.InputDataFields.groundtruth_area:
input_dict[fields.InputDataFields.groundtruth_area],
fields.InputDataFields.groundtruth_is_crowd:
input_dict[fields.InputDataFields.groundtruth_is_crowd],
fields.InputDataFields.groundtruth_difficult:
tf.cast(input_dict[fields.InputDataFields.groundtruth_difficult],
tf.int32)
}
if fields.InputDataFields.groundtruth_instance_masks in input_dict:
labels[fields.InputDataFields.
groundtruth_instance_masks] = input_dict[
fields.InputDataFields.groundtruth_instance_masks]
# Add a batch dimension to the tensors.
features = {
key: tf.expand_dims(features[key], axis=0)
for key, feature in features.items()
}
labels = {
key: tf.expand_dims(labels[key], axis=0)
for key, label in labels.items()
}
return features, labels
def get_next(config):
return dataset_builder.make_initializable_iterator(
dataset_builder.build(config)).get_next()
def graph_fn_last_batch():
dataset = dataset_builder.build(input_reader_proto,
batch_size=8,
input_context=input_context)
dataset = dataset.skip(8)
return get_iterator_next_for_testing(dataset, self.is_tf2())
def graph_fn():
dataset = dataset_builder.build(input_reader_proto,
batch_size=8)
dataset = dataset.skip(i)
return get_iterator_next_for_testing(dataset, self.is_tf2())
def graph_fn():
return dataset_builder.make_initializable_iterator(
dataset_builder.build(
input_reader_proto,
transform_input_data_fn=one_hot_class_encoding_fn,
batch_size=2)).get_next()
def graph_fn():
return get_iterator_next_for_testing(
dataset_builder.build(input_reader_proto, batch_size=1),
self.is_tf2())
def main(_):
assert FLAGS.train_dir, '`train_dir` is missing.'
if FLAGS.task == 0: tf.gfile.MakeDirs(FLAGS.train_dir)
if FLAGS.pipeline_config_path:
configs = config_util.get_configs_from_pipeline_file(
FLAGS.pipeline_config_path)
if FLAGS.task == 0:
tf.gfile.Copy(FLAGS.pipeline_config_path,
os.path.join(FLAGS.train_dir, 'pipeline.config'),
overwrite=True)
else:
configs = config_util.get_configs_from_multiple_files(
model_config_path=FLAGS.model_config_path,
train_config_path=FLAGS.train_config_path,
train_input_config_path=FLAGS.input_config_path)
if FLAGS.task == 0:
for name, config in [('model.config', FLAGS.model_config_path),
('train.config', FLAGS.train_config_path),
('input.config', FLAGS.input_config_path)]:
tf.gfile.Copy(config, os.path.join(FLAGS.train_dir, name),
overwrite=True)
model_config = configs['model']
train_config = configs['train_config']
input_config = configs['train_input_config']
model_fn = functools.partial(
model_builder.build,
model_config=model_config,
is_training=True)
#iterator = dataset_util.make_initializable_iterator(dataset_builder.build(input_config))
datasetmy = dataset_builder.build(input_config)
iterator = datasetmy.make_initializable_iterator()
def get_next(config):
return iterator.get_next()
create_input_dict_fn = functools.partial(get_next, input_config)
data_augmentation_options = [
preprocessor_builder.build(step)
for step in train_config.data_augmentation_options]
input_queue = trainer.create_input_queue(
train_config.batch_size, create_input_dict_fn,
train_config.batch_queue_capacity,
train_config.num_batch_queue_threads,
train_config.prefetch_queue_capacity, data_augmentation_options)
tensors = input_queue.dequeue()
#print all tensors in tfrecord
print(tensors)
groundtruth_difficult = tensors[0]['groundtruth_difficult']
groundtruth_group_of = tensors[0]['groundtruth_group_of']
groundtruth_weights = tensors[0]['groundtruth_weights']
groundtruth_is_crowd = tensors[0]['groundtruth_is_crowd']
key = tensors[0]['key']
groundtruth_boxes = tensors[0]['groundtruth_boxes']
image = tensors[0]['image']
groundtruth_area = tensors[0]['groundtruth_area']
groundtruth_classes = tensors[0]['groundtruth_classes']
filename = tensors[0]['filename']
num_groundtruth_boxes = tensors[0]['num_groundtruth_boxes']
source_id = tensors[0]['source_id']
init_op=tf.initialize_all_variables()
with tf.Session() as sess:
sess.run(iterator.initializer)
sess.run(tf.tables_initializer())
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(coord=coord)
sess.run(init_op)
for i in range(10):
groundtruth_weights_val,groundtruth_difficult_val,groundtruth_group_of_val,groundtruth_is_crowd_val,key_val,groundtruth_boxes_val,image_val,groundtruth_area_val,groundtruth_classes_val,filename_val,num_groundtruth_boxes_val,source_id_val = \
sess.run([groundtruth_weights,groundtruth_difficult,groundtruth_group_of,groundtruth_is_crowd,key,groundtruth_boxes,image,groundtruth_area,groundtruth_classes,filename,num_groundtruth_boxes,source_id])
# print(groundtruth_weights_val)
print(groundtruth_boxes_val)
# print(groundtruth_difficult_val)
# print(groundtruth_group_of_val)
# print(groundtruth_is_crowd_val)
# print(key_val)
# print(image_val)
# print(groundtruth_area_val)
print(groundtruth_classes_val)
print(filename_val)
print(num_groundtruth_boxes_val)
# print(source_id_val)
image_val = image_val[0]
image_val = image_val.astype(np.uint8)
# cv2.imshow('image', image_val)
# cv2.waitKey()
# plt.imshow(image_val)
# plt.show()
print('finish')
#plot bbox on image
plt.switch_backend("TkAgg")
classes_val = groundtruth_classes_val
boxes_val = groundtruth_boxes_val
scores_val = [1.0]*num_groundtruth_boxes_val
image_np = image_val
image_np_origin = image_val.copy()
NUM_CLASSES = 90
IMAGE_SIZE = (12, 8)
PATH_TO_LABELS = '../../data/mscoco_label_map.pbtxt'
label_map = label_map_util.load_labelmap(PATH_TO_LABELS)
categories = label_map_util.convert_label_map_to_categories(label_map, max_num_classes=NUM_CLASSES,
use_display_name=True)
category_index = label_map_util.create_category_index(categories)
vis_util.visualize_boxes_and_labels_on_image_array(
image_np,
boxes_val,
np.squeeze(classes_val).astype(np.int32),
np.squeeze(scores_val),
category_index,
use_normalized_coordinates=True,
line_thickness=8)
plt.figure(figsize=IMAGE_SIZE)
plt.subplot(121)
plt.imshow(image_np)
plt.subplot(122)
plt.imshow(image_np_origin)
plt.show()
print('finish')
pass
coord.request_stop()
coord.join(threads)
def _eval_input_fn(params=None):
"""Returns `features` and `labels` tensor dictionaries for evaluation.
Args:
params: Parameter dictionary passed from the estimator.
Returns:
features: Dictionary of feature tensors.
features[fields.InputDataFields.image] is a [1, H, W, C] float32 tensor
with preprocessed images.
features[HASH_KEY] is a [1] int32 tensor representing unique
identifiers for the images.
features[fields.InputDataFields.true_image_shape] is a [1, 3]
int32 tensor representing the true image shapes, as preprocessed
images could be padded.
features[fields.InputDataFields.original_image] is a [1, H', W', C]
float32 tensor with the original image.
labels: Dictionary of groundtruth tensors.
labels[fields.InputDataFields.groundtruth_boxes] is a [1, num_boxes, 4]
float32 tensor containing the corners of the groundtruth boxes.
labels[fields.InputDataFields.groundtruth_classes] is a
[num_boxes, num_classes] float32 one-hot tensor of classes.
labels[fields.InputDataFields.groundtruth_area] is a [1, num_boxes]
float32 tensor containing object areas.
labels[fields.InputDataFields.groundtruth_is_crowd] is a [1, num_boxes]
bool tensor indicating if the boxes enclose a crowd.
labels[fields.InputDataFields.groundtruth_difficult] is a [1, num_boxes]
int32 tensor indicating if the boxes represent difficult instances.
-- Optional --
labels[fields.InputDataFields.groundtruth_instance_masks] is a
[1, num_boxes, H, W] float32 tensor containing only binary values,
which represent instance masks for objects.
Raises:
TypeError: if the `eval_config` or `eval_input_config` are not of the
correct type.
"""
del params
if not isinstance(eval_config, eval_pb2.EvalConfig):
raise TypeError('For eval mode, the `eval_config` must be a '
'train_pb2.EvalConfig.')
if not isinstance(eval_input_config, input_reader_pb2.InputReader):
raise TypeError('The `eval_input_config` must be a '
'input_reader_pb2.InputReader.')
if not isinstance(model_config, model_pb2.DetectionModel):
raise TypeError('The `model_config` must be a '
'model_pb2.DetectionModel.')
num_classes = config_util.get_number_of_classes(model_config)
model = model_builder.build(model_config, is_training=False)
image_resizer_config = config_util.get_image_resizer_config(model_config)
image_resizer_fn = image_resizer_builder.build(image_resizer_config)
transform_data_fn = functools.partial(
transform_input_data, model_preprocess_fn=model.preprocess,
image_resizer_fn=image_resizer_fn,
num_classes=num_classes,
data_augmentation_fn=None,
retain_original_image=True)
dataset = dataset_builder.build(eval_input_config,
transform_input_data_fn=transform_data_fn)
input_dict = dataset_util.make_initializable_iterator(dataset).get_next()
hash_from_source_id = tf.string_to_hash_bucket_fast(
input_dict[fields.InputDataFields.source_id], HASH_BINS)
features = {
fields.InputDataFields.image:
input_dict[fields.InputDataFields.image],
fields.InputDataFields.original_image:
input_dict[fields.InputDataFields.original_image],
HASH_KEY: tf.cast(hash_from_source_id, tf.int32),
fields.InputDataFields.true_image_shape:
input_dict[fields.InputDataFields.true_image_shape]
}
labels = {
fields.InputDataFields.groundtruth_boxes:
input_dict[fields.InputDataFields.groundtruth_boxes],
fields.InputDataFields.groundtruth_classes:
input_dict[fields.InputDataFields.groundtruth_classes],
fields.InputDataFields.groundtruth_area:
input_dict[fields.InputDataFields.groundtruth_area],
fields.InputDataFields.groundtruth_is_crowd:
input_dict[fields.InputDataFields.groundtruth_is_crowd],
fields.InputDataFields.groundtruth_difficult:
tf.cast(input_dict[fields.InputDataFields.groundtruth_difficult],
tf.int32)
}
if fields.InputDataFields.groundtruth_instance_masks in input_dict:
labels[fields.InputDataFields.groundtruth_instance_masks] = input_dict[
fields.InputDataFields.groundtruth_instance_masks]
# Add a batch dimension to the tensors.
features = {
key: tf.expand_dims(features[key], axis=0)
for key, feature in features.items()
}
labels = {
key: tf.expand_dims(labels[key], axis=0)
for key, label in labels.items()
}
return features, labels
def get_next(config):
return dataset_util.make_initializable_iterator(
dataset_builder.build(config)).get_next()
def _train_input_fn(params=None):
"""Returns `features` and `labels` tensor dictionaries for training.
Args:
params: Parameter dictionary passed from the estimator.
Returns:
features: Dictionary of feature tensors.
features[fields.InputDataFields.image] is a [batch_size, H, W, C]
float32 tensor with preprocessed images.
features[HASH_KEY] is a [batch_size] int32 tensor representing unique
identifiers for the images.
features[fields.InputDataFields.true_image_shape] is a [batch_size, 3]
int32 tensor representing the true image shapes, as preprocessed
images could be padded.
labels: Dictionary of groundtruth tensors.
labels[fields.InputDataFields.num_groundtruth_boxes] is a [batch_size]
int32 tensor indicating the number of groundtruth boxes.
labels[fields.InputDataFields.groundtruth_boxes] is a
[batch_size, num_boxes, 4] float32 tensor containing the corners of
the groundtruth boxes.
labels[fields.InputDataFields.groundtruth_classes] is a
[batch_size, num_boxes, num_classes] float32 one-hot tensor of
classes.
labels[fields.InputDataFields.groundtruth_weights] is a
[batch_size, num_boxes] float32 tensor containing groundtruth weights
for the boxes.
-- Optional --
labels[fields.InputDataFields.groundtruth_instance_masks] is a
[batch_size, num_boxes, H, W] float32 tensor containing only binary
values, which represent instance masks for objects.
labels[fields.InputDataFields.groundtruth_keypoints] is a
[batch_size, num_boxes, num_keypoints, 2] float32 tensor containing
keypoints for each box.
Raises:
TypeError: if the `train_config` or `train_input_config` are not of the
correct type.
"""
if not isinstance(train_config, train_pb2.TrainConfig):
raise TypeError('For training mode, the `train_config` must be a '
'train_pb2.TrainConfig.')
if not isinstance(train_input_config, input_reader_pb2.InputReader):
raise TypeError('The `train_input_config` must be a '
'input_reader_pb2.InputReader.')
if not isinstance(model_config, model_pb2.DetectionModel):
raise TypeError('The `model_config` must be a '
'model_pb2.DetectionModel.')
data_augmentation_options = [
preprocessor_builder.build(step)
for step in train_config.data_augmentation_options
]
data_augmentation_fn = functools.partial(
augment_input_data,
data_augmentation_options=data_augmentation_options)
model = model_builder.build(model_config, is_training=True)
image_resizer_config = config_util.get_image_resizer_config(
model_config)
image_resizer_fn = image_resizer_builder.build(image_resizer_config)
transform_data_fn = functools.partial(
transform_input_data,
model_preprocess_fn=model.preprocess,
image_resizer_fn=image_resizer_fn,
num_classes=config_util.get_number_of_classes(model_config),
data_augmentation_fn=data_augmentation_fn)
dataset = dataset_builder.build(
train_input_config,
transform_input_data_fn=transform_data_fn,
batch_size=params['batch_size']
if params else train_config.batch_size,
max_num_boxes=train_config.max_number_of_boxes,
num_classes=config_util.get_number_of_classes(model_config),
spatial_image_shape=config_util.get_spatial_image_size(
image_resizer_config))
tensor_dict = dataset_util.make_initializable_iterator(
dataset).get_next()
hash_from_source_id = tf.string_to_hash_bucket_fast(
tensor_dict[fields.InputDataFields.source_id], HASH_BINS)
features = {
fields.InputDataFields.image:
tensor_dict[fields.InputDataFields.image],
HASH_KEY:
tf.cast(hash_from_source_id, tf.int32),
fields.InputDataFields.true_image_shape:
tensor_dict[fields.InputDataFields.true_image_shape]
}
labels = {
fields.InputDataFields.num_groundtruth_boxes:
tensor_dict[fields.InputDataFields.num_groundtruth_boxes],
fields.InputDataFields.groundtruth_boxes:
tensor_dict[fields.InputDataFields.groundtruth_boxes],
fields.InputDataFields.groundtruth_classes:
tensor_dict[fields.InputDataFields.groundtruth_classes],
fields.InputDataFields.groundtruth_weights:
tensor_dict[fields.InputDataFields.groundtruth_weights]
}
if fields.InputDataFields.groundtruth_keypoints in tensor_dict:
labels[fields.InputDataFields.groundtruth_keypoints] = tensor_dict[
fields.InputDataFields.groundtruth_keypoints]
if fields.InputDataFields.groundtruth_instance_masks in tensor_dict:
labels[fields.InputDataFields.
groundtruth_instance_masks] = tensor_dict[
fields.InputDataFields.groundtruth_instance_masks]
return features, labels
