Update a variable with tf.while_loop in Tensorflow

Question

I want to update a variable in Tensorflow and for that reason I use the tf.while_loop like:

a = tf.Variable([0, 0, 0, 0, 0, 0] , dtype = np.int16)

i = tf.consta         


        

Answer 1

This isn't obvious until one codes and executes. It is like this pattern

import tensorflow as tf


def cond(size, i):
    return tf.less(i,size)

def body(size, i):

    a = tf.get_variable("a",[6],dtype=tf.int32,initializer=tf.constant_initializer(0))
    a = tf.scatter_update(a,i,i)

    tf.get_variable_scope().reuse_variables() # Reuse variables
    with tf.control_dependencies([a]):
        return (size, i+1)

with tf.Session() as sess:

    i = tf.constant(0)
    size = tf.constant(6)
    _,i = tf.while_loop(cond,
                    body,
                    [size, i])

    a = tf.get_variable("a",[6],dtype=tf.int32)

    init = tf.initialize_all_variables()
    sess.run(init)

    print(sess.run([a,i]))

Output is

[array([0, 1, 2, 3, 4, 5]), 6]

1. tf.get_variableGets an existing variable with these parameters or create a new one.
2. tf.control_dependencies It is a happens-before relationship. In this case I understand that the scatter_update happens before the while increments and returns. It doesn't update without this.

Note : I didn't really understand the meaning or cause of the error. I get that too.