Ligang Yan颜力刚

Tag

numpy

1 post

  1. Gradient Descent, Line by Line: Twelve Lines of Python and the Calculus Behind ThemA twelve-line function fits the best straight line through a cloud of points: it starts flat at zero and 900 times over asks which way is less wrong. Taking it apart one line at a time — what a gradient actually is, where the formulas come from, why the minus sign, and what alpha does.