How can I run a loop with a tensor as its range? (in tensorflow)

Question

I want to have a for loop that the number of its iterations is depend on a tensor value. For example:

for i in tf.range(input_placeholder[1,1]):
  # do something         


        

Answer 1

The type of the return value of TensorFlow Python API functions, including tf.range is a Tensor. A Tensor is a symbolic handle to node in a graph that represents computation. You perform the actual computation by calling the eval method on a Tensor, or by passing the object to run method of a Session. In your case, perhaps what you intended to do was simply iterate over numpy's range.

for in in np.range(...):
  # do something

Answer 2

To do this you will need to use the tensorflow while loop (tf.while_loop) as follows:

i = tf.constant(0)
while_condition = lambda i: tf.less(i, input_placeholder[1, 1])
def body(i):
    # do something here which you want to do in your loop
    # increment i
    return [tf.add(i, 1)]

# do the loop:
r = tf.while_loop(while_condition, body, [i])