ToDO Graphs for every loss function and it’s derivative reflow gradient descent document https://github.com/kennethleungty/MLOps-Specialization-Notes