CODEWITHSUNDEEP
tensorflowlstm
A Sharma
A Sharma

Reputation: 27

tf.train.batch is showing insufficient elements

I am working on LSTM based models. the data consists of 80000 images. I am using batchsize of 1 and getting the following as error log:

OutOfRangeError (see above for traceback): PaddingFIFOQueue '_1_Train_data/batch/padding_fifo_queue' is closed and has insufficient elements (requested 1, current size 0) [[Node: Train_data/batch = QueueDequeueManyV2[component_types=[DT_FLOAT, DT_STRING, DT_INT32], timeout_ms=-1, _device="/job:localhost/replica:0/task:0/cpu:0"](Train_data/batch/padding_fifo_queue, Train_data/batch/n)]]

Can someone suggests what can be the possible issue? as the fifoqueue size is shown as 0 for all possible batch sizes I tried.

Upvotes: 0

Views: 897

Answers (2)

A Sharma
A Sharma

Reputation: 27

some of the images were corrupted in the database which was causing the program to run into this error. Removed those images and now working fine.

Upvotes: 0

Mohamed
Mohamed

Reputation: 486

The error doesn't have to do with LSTM, you are getting it from tf.train.batch. You have to initialize your TF local variable along with global variables. From this open issue https://github.com/tensorflow/tensorflow/issues/1045, it seems that the order of initialization matters. 

global_init_op = tf.global_variables_initializer()
local_init_op = tf.local_variables_initializer()

with tf.Session() as sess:
    sess.run(global_init_op)
    sess.run(local_init_op)
    # rest of your code

Upvotes: 1

Related Questions