No Grad

no_grad() is tensor state function that controls whehter gradient chain graphs are created.

use with ntk.no_grad(): to enter no gradient mode. No gradient mode is meant to be used post training when tracking gradients is not necessray.

with ntk.no_grad():
    model(data)