W2A1 "Wrong values when training=True."

I double, triple check my code, and still getting this error. I set breakpoint and check the shape and value of X an X_shortcut, everything seem correct to me.

Not sure what else missing

Link to assignment Residual Networks | Coursera

@psyquy can you post a screenshot of the errors you are getting ?

I update the question with screenshot, please check

The most common mistake in this Exercise is to input the X for the SHORTCUT PATH which is wrong. It shouldn’t be X.

i did use X_shortcut for shortcut calculation,

X_shortcut = X 
X_shortcut = <thelayer>()(X_shortcut)
X =Add()([X, X_shortcut])

After restart a few time (without changing any code). now it pass. :thinking:

i download to my local machine, and it failed with the same error again

Glad to know that…

it seems to me that the test suite is not deterministic. report issue to coursera