Question

我想在TensorFlow上构建一个CNN模型框架。但我一直被困在负载数据中。我的数据集是MIT-300，我想输入6张图片作为数据集，他们将通过网络，最后每张图片将与显着性图片（地面真实图片）进行比较。训练数据和基本事实都是'.jpg'的形式。

我试着像下面这样加载，但一直收到错误。现在我被告知这样：

OutOfRangeError (see above for traceback): FIFOQueue '_1_batch/fifo_queue' is closed and has insufficient elements (requested 3, current size 0)
     [[Node: batch = QueueDequeueManyV2[component_types=[DT_FLOAT, DT_FLOAT], timeout_ms=-1, _device="/job:localhost/replica:0/task:0/cpu:0"](batch/fifo_queue, batch/n)]]

Mycodes如下：

from __future__ import print_function
from __future__ import print_function
import tensorflow as tf
import os

IMAGE_WIDTH = 1024
IMAGE_HEIGHT = 768
NUMBER_OF_CHANNELS = 3
SOURCE_DIR = '/home/lvjc2010/Documents/mnist/data/'
TRAINING_IMAGES_DIR = SOURCE_DIR + 'train/'
LIST_FILE_NAME = 'train.txt'
BATCH_SIZE = 3
TRAINING_SET_SIZE = 6


def create_photo_and_label_batches(source_directory):
    # read the list of photo IDs and labels
    photos_list = open(source_directory + LIST_FILE_NAME, 'r')
    filenames_list = []
    labels_list = []
    # get lists of photo file names and labels
    for line in photos_list:
        filenames_list.append(source_directory + line.split(' ')[0])
        labels_list.append(source_directory + line.split(' ')[1])
    # convert the lists to tensors
    filenames = tf.convert_to_tensor(filenames_list, dtype=tf.string)
    labels = tf.convert_to_tensor(labels_list, dtype=tf.string)
    # create queue with filenames and labels
    file_names_queue, labels_names_queue =\
        tf.train.slice_input_producer([filenames, labels], shuffle=False)
    ## IT WORKED


    # convert filenames of photos to input vectors
    photos_queue = tf.read_file(file_names_queue)  # convert filenames to content
    photos_queue = tf.image.decode_image(photos_queue, channels=NUMBER_OF_CHANNELS)
    photos_queue.set_shape([IMAGE_HEIGHT, IMAGE_WIDTH, NUMBER_OF_CHANNELS])
    photos_queue = tf.cast(photos_queue, tf.float32)*(1./255)-0.5  # convert uint8 to float32
    photos_queue = tf.reshape(photos_queue, [-1])  # flatten the tensor

    labels_queue = tf.read_file(labels_names_queue)  # convert filenames to content
    labels_queue = tf.image.decode_jpeg(labels_queue, channels=NUMBER_OF_CHANNELS)
    labels_queue.set_shape([IMAGE_HEIGHT, IMAGE_WIDTH, NUMBER_OF_CHANNELS])
    labels_queue = tf.to_float(labels_queue)  # convert uint8 to float32
    labels_queue = tf.reshape(labels_queue, [-1])  # flatten the tensor

    # slice the data into mini batches
    return tf.train.batch([photos_queue, labels_queue], batch_size=BATCH_SIZE)
    # return tf.train.batch(photos_queue, batch_size=BATCH_SIZE)

def weight_variable(shape):
    """weight_variable generates a weight variable of a given shape."""
    initial = tf.truncated_normal(shape, stddev=0.1)
    return tf.Variable(initial)


def bias_variable(shape):
    """bias_variable generates a bias variable of a given shape."""
    initial = tf.constant(0.1, shape=shape)
    return tf.Variable(initial)


if __name__ == '__main__':
  # load the training data
  training_photo_batch, training_label_batch =\
      create_photo_and_label_batches(TRAINING_IMAGES_DIR)
  # with tf.Session() as sess:
  #     print (sess.run(training_photo_batch))
  # create the model
  x = training_photo_batch
  # W = tf.Variable(tf.zeros([IMAGE_WIDTH * IMAGE_HEIGHT * NUMBER_OF_CHANNELS, 10],
  #    dtype=tf.float32))  # weights tensor
  b = tf.Variable(tf.zeros(IMAGE_WIDTH * IMAGE_HEIGHT * NUMBER_OF_CHANNELS, dtype=tf.float32))  # bias
  y_ = training_label_batch
  # y = tf.matmul(x, W) + b
  y = tf.add(x,b)


  # define loss and optimizer

  cross_entropy = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=y, labels=y_))
  train_step = tf.train.GradientDescentOptimizer(0.5).minimize(cross_entropy)


  # do the training
  sess = tf.InteractiveSession()
  config = tf.ConfigProto()
  config.gpu_options.allow_growth = True
  tf.initialize_all_variables().run()
  coord = tf.train.Coordinator()
  threads = tf.train.start_queue_runners(coord=coord)



  # with tf.name_scope('accuracy'):
  #   correct_prediction = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))
  #   correct_prediction = tf.cast(correct_prediction, tf.float32)
  # accuracy = tf.reduce_mean(correct_prediction)



  for i in range(TRAINING_SET_SIZE // BATCH_SIZE):
    print('hello')
    sess.run(train_step)
    # train_accuracy = accuracy.eval(feed_dict={
    #     x: y, y_: y_})
  # print('step %d, training accuracy %g' % (i, train_accuracy))

  # stop the queue threads and properly close the session
  coord.request_stop()
  coord.join(threads)
  sess.close()

我使用sess.run在这里调试。我发现地面真相图像无法加载。

labels_queue = tf.read_file(labels_names_queue)  # convert filenames to content
    with tf.Session() as sess:
        coord = tf.train.Coordinator()
        threads = tf.train.start_queue_runners(sess=sess, coord=coord)
        k = sess.run(labels_queue)
        print('k', k)
    labels_queue = tf.image.decode_jpeg(labels_queue, channels=NUMBER_OF_CHANNELS)

它显示：

NotFoundError (see above for traceback): /home/lvjc2010/Documents/mnist/data/train/i05june05_static_street_boston_p1010764_fixMap.jpg

     [[Node: ReadFile_1 = ReadFile[_device="/job:localhost/replica:0/task:0/cpu:0"](input_producer/Gather_1)]]

为什么在我想加载数据集时收到OurOfRangeError？

0 个答案: