I don’t know if fine tuning works. But if it doesn’t, then are we assuming the underlying weights are optimal? At what point do we determine that a network is p

I don’t know if fine tuning works. But if it doesn’t, then are we assuming the underlying weights are optimal? At what point do we determine that a network is properly “trained” and any subsequent training is “fine tuning”.