This question is in reference to http://stackoverflow.com/a/43662693/4571192.
I am also facing the same issue,but was not able to understand the solution pointed out. where should this be added:images = tf.identity(images, name='Inputs') And how does impact the network and how is it linked to the previously retrained network.
Kindly help in resolving this error.
i did not add any comment there due to low reputation points.