tfrecords数据的读写

tfrecords_write.py

import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
import numpy as np

#定义一个writer->转化数据的格式->包装好一个example->将example序列化->写入

#将数据转化成相应的格式
def _int64_feature(value):
    return tf.train.Feature(int64_list=tf.train.Int64List(value=[value]))

def _bytes_feature(value):
    return tf.train.Feature(bytes_list=tf.train.BytesList(value=[value]))

mnist = input_data.read_data_sets('../MNIST_data',one_hot=True,dtype=tf.uint8)

images = mnist.train.images
labels = mnist.train.labels

pixels = images.shape[1]

num_examples = mnist.train.num_examples

filename = 'output.tfrecords'

writer = tf.python_io.TFRecordWriter(filename)

for i in range(num_examples):
    image_raw = images[i].tostring()
    example = tf.train.Example(features=tf.train.Features(feature={
        'pixels':_int64_feature(pixels),
        'label':_int64_feature(np.argmax(labels[i])),
        'image_raw':_bytes_feature(image_raw)
    }))
    writer.write(example.SerializeToString())

writer.close()

tfrecords_read.py

import tensorflow as tf

#定义reader->read一个example->解析->打印

reader = tf.TFRecordReader()

#使用队列来维护输入文件列表
filename_queue = tf.train.string_input_producer(['output.tfrecords'])

#获取序列化的数据,每次只读取一个
_,serialize_examples = reader.read(filename_queue)

#解析数据
features = tf.parse_single_example(
    serialize_examples,
    features={
        'image_raw':tf.FixedLenFeature([],tf.string),
        'pixels':tf.FixedLenFeature([],tf.int64),
        'label':tf.FixedLenFeature([],tf.int64)
    }
)
#将字符串解析成像素数组
image = tf.decode_raw(features['image_raw'],tf.uint8)
label = tf.cast(features['label'],tf.int32)
pixels = tf.cast(features['pixels'],tf.int32)

with tf.Session() as sess:
    #启动多线程
    threads = tf.train.start_queue_runners(sess,tf.train.Coordinator())

    for i in range(10):
        #每次都随机读取一个样例
        print(sess.run([label]))