Numpy 到 TFrecords:是否有更简单的方法来处理来自 tfrecords 的批量输入?
Numpy to TFrecords: Is there a more simple way to handle batch inputs from tfrecords?
我的问题是关于如何从多个(或分片的)tfrecords 获取批量输入。我已经阅读了示例 https://github.com/tensorflow/models/blob/master/inception/inception/image_processing.py#L410。基本流程是,以训练集为例,(1)首先生成一系列的tfrecords(例如,train-000-of-005
,train-001-of-005
,...),(2)从这些文件名中,生成一个列表并将它们送入 tf.train.string_input_producer
以获得队列,(3)同时生成一个 tf.RandomShuffleQueue
来做其他事情,(4)使用 tf.train.batch_join
生成批输入。
我觉得这很复杂,我不确定这个过程的逻辑。在我的例子中,我有一个 .npy
文件列表,我想生成分片 tfrecords(多个分离的 tfrecords,而不仅仅是一个大文件)。这些 .npy
个文件中的每一个都包含不同数量的正样本和负样本 (2 类)。一种基本方法是生成一个单一的大型 tfrecord 文件。但是文件太大 (~20Gb
)。所以我求助于分片 tfrecords。有没有更简单的方法来做到这一点?谢谢。
使用Dataset API
简化了整个过程。这是两个部分:(1): Convert numpy array to tfrecords
和 (2,3,4): read the tfrecords to generate batches
.
1。 从 numpy 数组创建 tfrecords:
def npy_to_tfrecords(...):
# write records to a tfrecords file
writer = tf.python_io.TFRecordWriter(output_file)
# Loop through all the features you want to write
for ... :
let say X is of np.array([[...][...]])
let say y is of np.array[[0/1]]
# Feature contains a map of string to feature proto objects
feature = {}
feature['X'] = tf.train.Feature(float_list=tf.train.FloatList(value=X.flatten()))
feature['y'] = tf.train.Feature(int64_list=tf.train.Int64List(value=y))
# Construct the Example proto object
example = tf.train.Example(features=tf.train.Features(feature=feature))
# Serialize the example to a string
serialized = example.SerializeToString()
# write the serialized objec to the disk
writer.write(serialized)
writer.close()
2。 使用数据集读取 tfrecords API (tensorflow >=1.2):
# Creates a dataset that reads all of the examples from filenames.
filenames = ["file1.tfrecord", "file2.tfrecord", ..."fileN.tfrecord"]
dataset = tf.contrib.data.TFRecordDataset(filenames)
# for version 1.5 and above use tf.data.TFRecordDataset
# example proto decode
def _parse_function(example_proto):
keys_to_features = {'X':tf.FixedLenFeature((shape_of_npy_array), tf.float32),
'y': tf.FixedLenFeature((), tf.int64, default_value=0)}
parsed_features = tf.parse_single_example(example_proto, keys_to_features)
return parsed_features['X'], parsed_features['y']
# Parse the record into tensors.
dataset = dataset.map(_parse_function)
# Shuffle the dataset
dataset = dataset.shuffle(buffer_size=10000)
# Repeat the input indefinitly
dataset = dataset.repeat()
# Generate batches
dataset = dataset.batch(batch_size)
# Create a one-shot iterator
iterator = dataset.make_one_shot_iterator()
# Get batch X and y
X, y = iterator.get_next()
我的问题是关于如何从多个(或分片的)tfrecords 获取批量输入。我已经阅读了示例 https://github.com/tensorflow/models/blob/master/inception/inception/image_processing.py#L410。基本流程是,以训练集为例,(1)首先生成一系列的tfrecords(例如,train-000-of-005
,train-001-of-005
,...),(2)从这些文件名中,生成一个列表并将它们送入 tf.train.string_input_producer
以获得队列,(3)同时生成一个 tf.RandomShuffleQueue
来做其他事情,(4)使用 tf.train.batch_join
生成批输入。
我觉得这很复杂,我不确定这个过程的逻辑。在我的例子中,我有一个 .npy
文件列表,我想生成分片 tfrecords(多个分离的 tfrecords,而不仅仅是一个大文件)。这些 .npy
个文件中的每一个都包含不同数量的正样本和负样本 (2 类)。一种基本方法是生成一个单一的大型 tfrecord 文件。但是文件太大 (~20Gb
)。所以我求助于分片 tfrecords。有没有更简单的方法来做到这一点?谢谢。
使用Dataset API
简化了整个过程。这是两个部分:(1): Convert numpy array to tfrecords
和 (2,3,4): read the tfrecords to generate batches
.
1。 从 numpy 数组创建 tfrecords:
def npy_to_tfrecords(...):
# write records to a tfrecords file
writer = tf.python_io.TFRecordWriter(output_file)
# Loop through all the features you want to write
for ... :
let say X is of np.array([[...][...]])
let say y is of np.array[[0/1]]
# Feature contains a map of string to feature proto objects
feature = {}
feature['X'] = tf.train.Feature(float_list=tf.train.FloatList(value=X.flatten()))
feature['y'] = tf.train.Feature(int64_list=tf.train.Int64List(value=y))
# Construct the Example proto object
example = tf.train.Example(features=tf.train.Features(feature=feature))
# Serialize the example to a string
serialized = example.SerializeToString()
# write the serialized objec to the disk
writer.write(serialized)
writer.close()
2。 使用数据集读取 tfrecords API (tensorflow >=1.2):
# Creates a dataset that reads all of the examples from filenames.
filenames = ["file1.tfrecord", "file2.tfrecord", ..."fileN.tfrecord"]
dataset = tf.contrib.data.TFRecordDataset(filenames)
# for version 1.5 and above use tf.data.TFRecordDataset
# example proto decode
def _parse_function(example_proto):
keys_to_features = {'X':tf.FixedLenFeature((shape_of_npy_array), tf.float32),
'y': tf.FixedLenFeature((), tf.int64, default_value=0)}
parsed_features = tf.parse_single_example(example_proto, keys_to_features)
return parsed_features['X'], parsed_features['y']
# Parse the record into tensors.
dataset = dataset.map(_parse_function)
# Shuffle the dataset
dataset = dataset.shuffle(buffer_size=10000)
# Repeat the input indefinitly
dataset = dataset.repeat()
# Generate batches
dataset = dataset.batch(batch_size)
# Create a one-shot iterator
iterator = dataset.make_one_shot_iterator()
# Get batch X and y
X, y = iterator.get_next()