梯度

梯度是向量,每一维都是对应维度的偏导数

In multiple dimensions, the gradient is the vector of (partial derivatives) along each dimension

原文地址:https://www.cnblogs.com/war1111/p/13933528.html