Sayantan Sarkar

In our discussion @shresthamalik pointed out this graph: ``` W | v X-->Y-->*I--->Z | ^ |________| ``` Data edges: `X->Y`, `I->Z`, `X->I`, `W->Y`. Control edges (starred): `Y->*I`. `I` is identity...

I think most of Adam's review comments are addressed or incorporated. The two things remaining are: 1) T& vs T* for return. TF style prefers T* for return/changeable variables. references...

Ok, will make the T& -> T* changes. Also tagged you in slack

TESTNOW

TESTNOW

Fail message: ``` self = def test_tanhgrad_2d(self): y = constant_op.constant( self.generate_random_numbers(30, 1.0, 10.0), shape=[10, 3]) y_delta = constant_op.constant( self.generate_random_numbers(30, 0.0, 10.0), shape=[10, 3]) out = tanh_grad(y, y_delta) def run_test(sess): return...