In TensorFlow,what's the meaning of ":0" in a Variable's name?

Question

import tensorflow as tf with tf.device('/gpu:0'):     foo = tf.Variable(1, name='foo')     assert foo.name == "foo:0" with tf.device('/gpu:1'):     bar = tf.Variable(1, name='bar')     assert bar.name == "bar:0" 

The above code returns true.I use with tf.device here to illustrate that the ":0" doesn't mean the variable lie on the specific device.So what's the meaning of the ":0" in the variable's name(foo and bar in this example)?

Answer 1

It has to do with representation of tensors in underlying API. A tensor is a value associated with output of some op. In case of variables, there's a Variable op with one output. An op can have more than one output, so those tensors get referenced to as <op>:0, <op>:1 etc. For instance if you use tf.nn.top_k, there are two values created by this op, so you may see TopKV2:0 and TopKV2:1

a,b=tf.nn.top_k([1], 1) print a.name # => 'TopKV2:0' print b.name # => 'TopKV2:1' 

How to understand the term `tensor` in TensorFlow?