Run multiple pre-trained Tensorflow nets at the same time

Question

What I would like to do is to run multiple pre-trained Tensorflow nets at the same time. Because the names of some variables inside each net can be the same, the common solu

Answer 1

I have the same problem that bothered me a long time. I found a good solution here: Loading two models from Saver in the same Tensorflow session and TensorFlow checkpoint save and read.

The default behavior for a tf.train.Saver() is to associate each variable with the name of the corresponding op. This means that each time you construct a tf.train.Saver(), it includes all of the variables for the previous calls. Therefore, you should create different graphs and run different sessions with them.