machine learning - Tensorflow Training and Validation Input Queue Separation -

i tried replicate convolutional network results using tensorflow. used marvin teichmann's implementation github. need write training wrapper. create 2 graphs share variables , 2 input queues, 1 training , 1 validation. test training wrapper, used 2 short lists of training , validation files , validation after every training epoch. printed out shape of every image input queue check whether correct input. however, after started training, seems images training queue being dequeued. both training , validation graphs take input training queue , validation queue never accessed. can explain , solve problem?

here's part of relevant code:

def get_data(image_name_list, num_epochs, scope_name, num_class = num_class):     tf.variable_scope(scope_name) scope:         images_path = [os.path.join(dataset_dir, i+'.jpg') in image_name_list]         gts_path = [os.path.join(gt_dir, i+'.png') in image_name_list]         seed = random.randint(0, 2147483647)         image_name_queue = tf.train.string_input_producer(images_path, num_epochs=num_epochs, shuffle=false, seed = seed)         gt_name_queue = tf.train.string_input_producer(gts_path, num_epochs=num_epochs, shuffle=false, seed = seed)         reader = tf.wholefilereader()         image_key, image_value = reader.read(image_name_queue)         my_image = tf.image.decode_jpeg(image_value)         my_image = tf.cast(my_image, tf.float32)         my_image = tf.expand_dims(my_image, 0)         gt_key, gt_value = reader.read(gt_name_queue)         # gt stands ground truth         my_gt = tf.cast(tf.image.decode_png(gt_value, channels = 1), tf.float32)         my_gt = tf.one_hot(tf.cast(my_gt, tf.int32), num_class)         return my_image, my_gt  train_image, train_gt = get_data(train_files, num_epoch, 'training') val_image, val_gt = get_data(val_files, num_epoch, 'validation') tf.variable_scope('fcn16') scope:         train_vgg16_fcn = fcn16_vgg.fcn16vgg()         train_vgg16_fcn.build(train_image, train=true, num_classes=num_class, keep_prob = keep_prob)         scope.reuse_variables()         val_vgg16_fcn = fcn16_vgg.fcn16vgg()         val_vgg16_fcn.build(val_image, train=false, num_classes=num_class, keep_prob = 1) """ define loss, evaluation metric, summary, saver in computation graph. initialize variables , start session. """ epoch in range(starting_epoch, num_epoch):     in range(train_num):         _, loss_value, shape = sess.run([train_op, train_entropy_loss, tf.shape(train_image)])         print shape     in range(val_num):         loss_value, shape = sess.run([val_entropy_loss, tf.shape(val_image)])         print shape 

to make sure reading different images run:

[train_image_np, val_image_np] = sess.run([train_image, val_image]) 

to reuse variables better , safer:

with tf.variable_scope('fcn16') scope:    train_vgg16_fcn = fcn16_vgg.fcn16vgg()      train_vgg16_fcn.build(train_image, train=true, num_classes=num_class, keep_prob = keep_prob) tf.variable_scope(scope, reuse=true):    val_vgg16_fcn = fcn16_vgg.fcn16vgg()    val_vgg16_fcn.build(val_image, train=false, num_classes=num_class, keep_prob = 1)