How to switch between training and validation dataset with tf.MonitoredTrainingSession?

Question

I want to use feedable iterator design in tensorflow Dataset API, so I can switch to validation data after some training steps. But if I switched to validation data

Answer 1

I would suggest to catch tf.errors.OutOfRangeError raised at the end of the validation dataset (you can also check the processing multiple epochs section in the official API for another solution using the repeat dataset ):

while not sess.should_stop():
    x = sess.run(next_element, feed_dict={handle: training_handle})
    count_training += 1
    print('{} [training] {}'.format(count_training, x.shape))

    # we do periodic validation
    if count_training % 4 == 0:
        sess.run(validation_iterator.initializer)
        count_validation = 0
        while True:
            try:
                y = sess.run(next_element, feed_dict={handle: validation_handle})
                count_validation += 1
                print('  {} [validation] {}'.format(count_validation, y.shape))
            except tf.errors.OutOfRangeError:
                break

This piece of code prints:

1 [training] (4,)  
2 [training] (4,)  
3 [training] (4,)  
4 [training] (4,)  
  1 [validation] (4,)  
  2 [validation] (4,)  
5 [training] (4,)
6 [training] (4,)
7 [training] (4,)
8 [training] (4,)
  1 [validation] (4,)
  2 [validation] (4,)