fine tuning

(ML Spec)

fine tuning takes parameters initialized or gotten from supervised pre-training of a different task and then run gradient descent to fine tune weights to suit the specific task at hand.