HI,
I have a very simple Tensorflow graph:
input -> conv -> x1
input -> conv -> x2
x1 + x2 -> y
I use the same weight and other parameters in the conv so x1 and x2 is the same value, and therefore, y is twice of one of them.
mvNCCheck gives total wrong values for y, which is identical to x1/x2 itself without doing the adding!!
Any helps?
