def _parse_example_fn(example_proto):
"""
Parse the input `tf.Example` proto using the features_spec
Args:
example_proto: tfrecord Example protobuf data
Returns:
features: parsed features extracted from the protobuf
labels: parsed label extracted from the protobuf
"""
features = io.parse_single_example(serialized=example_proto, features=features_spec)
features_dict = dict()
# Process all features
for feature_info in feature_config.get_all_features():
feature_node_name = feature_info.get("node_name", feature_info["name"])
default_tensor = tf.constant(
value=feature_config.get_default_value(feature_info), dtype=feature_info["dtype"],
)
feature_tensor = features.get(feature_info["name"], default_tensor)
feature_tensor = tf.expand_dims(feature_tensor, axis=0)
feature_tensor = preprocess_feature(feature_tensor, feature_info, preprocessing_map)
features_dict[feature_node_name] = feature_tensor
labels = features_dict.pop(feature_config.get_label(key="name"))
return features_dict, labels
def decode(self, serialized_example):
""" Load final output to neural network training procedure
:param serialized_example: compressed example recorded
on TFRecordWriter
:return: image, label as one hot encoder
"""
features = io.parse_single_example(serialized_example,
self.get_template())
image = io.decode_raw(features['image'], 'uint8')
image = reshape(image, self.shape + (3, ))
image = cast(image, dtype='float32') / 255
return image, one_hot(features['label'], self.labels)
def extract_features_from_proto(self, serialized):
"""
Parse the serialized proto string to extract features
Parameters
----------
proto: tf.Tensor
A scalar string tensor that is the serialized form of a TFRecord object
Returns
-------
dict of Tensors
Dictionary of features extracted from the proto as per the features_spec
"""
return io.parse_single_example(serialized=serialized, features=self.features_spec)
def _parse_example_fn(example_proto):
"""
Parse the input `tf.Example` proto using the features_spec
Parameters
----------
example_proto : string
serialized tfrecord Example protobuf message
Returns
-------
features : dict
parsed features as `tf.Tensor` objects extracted from the protobuf
labels : `tf.Tensor`
parsed label as a `tf.Tensor` object extracted from the protobuf
"""
features = io.parse_single_example(serialized=example_proto, features=features_spec)
features_dict = dict()
# Process all features, including label.
for feature_info in feature_config.get_all_features():
feature_node_name = feature_info.get("node_name", feature_info["name"])
default_tensor = tf.constant(
value=feature_config.get_default_value(feature_info), dtype=feature_info["dtype"],
)
feature_tensor = features.get(feature_info["name"], default_tensor)
feature_tensor = tf.expand_dims(feature_tensor, axis=0)
feature_tensor = preprocess_feature(feature_tensor, feature_info, preprocessing_map)
features_dict[feature_node_name] = feature_tensor
labels = features_dict.pop(feature_config.get_label(key="name"))
return features_dict, labels