TensorFlow 实战Google深度学习框架勘误

《TensorFlow:实战Google深度学习框架》7.3.2 输入文件队列章节中多线程读取记录的代码如下:

import tensorflow as tf

files = tf.train.match_filenames_once("/tmp/data.tfrecords-*")
filename_queue = tf.train.string_input_producer(files, shuffle=False)

reader = tf.TFRecordReader()
_, serialized_example = reader.read(filename_queue)
features = tf.parse_single_example(serialized_example,
                features={
                    'i': tf.FixedLenFeature([], tf.int64),
                    'j': tf.FixedLenFeature([], tf.int64),
                })
with tf.Session() as sess:
    tf.global_variables_initializer().run()
   
    print([str(i.name) for i in tf.local_variables()])

    print(sess.run(files))

    coord = tf.train.Coordinator()
    threads = tf.train.start_queue_runners(sess=sess, coord=coord)

    for i in range(6):
        print(sess.run([features['i'], features['j']]))

    coord.request_stop()
    coord.join(threads)

运行上述代码会报如下错误:

tensorflow.python.framework.errors_impl.FailedPreconditionError: Attempting to use uninitialized value matching_filenames
         [[Node: _retval_matching_filenames_0_0 = _Retval[T=DT_STRING, index=0, _device="/job:localhost/replica:0/task:0/device:CPU:0"](matching_filenames)]]

很诡异的代码中好像没有matching_filenames这个本地变量,但是如果看一下match_filenames_once这个函数的话就会发现问题所在了:

@tf_export("train.match_filenames_once")
def match_filenames_once(pattern, name=None):
  """Save the list of files matching pattern, so it is only computed once.

  NOTE: The order of the files returned can be non-deterministic.

  Args:
    pattern: A file pattern (glob), or 1D tensor of file patterns.
    name: A name for the operations (optional).

  Returns:
    A variable that is initialized to the list of files matching the pattern(s).
  """
  with ops.name_scope(name, "matching_filenames", [pattern]) as name:
    return vs.variable(
        name=name, initial_value=io_ops.matching_files(pattern),
        trainable=False, validate_shape=False,
        collections=[ops.GraphKeys.LOCAL_VARIABLES])

所以解决方法很简单:

...
with tf.Session() as sess:
    tf.global_variables_initializer().run()
    tf.local_variables_initializer().run()
...