cr:http://blog.csdn.net/txwh0820/article/details/46392293 一.矩阵的迹求导法则 1. 复杂矩阵问题求导方法:可以从小到大,从scalar到vector再到matrix 2. x is a column vector, A is a matrix d(A∗x)/dx=A d(xT∗A)/dxT=A d(xT∗A)/dx=AT d(xT∗A∗x)/dx=xT(AT+A) 3. Practice: 4. 矩阵求导计算法则 求导公式(撇号为
转载自: http://blog.csdn.net/txwh0820/article/details/46392293 矩阵的迹求导法则 1. 复杂矩阵问题求导方法:可以从小到大,从scalar到vector再到matrix 2. x is a column vector, A is a matrix d(A∗x)/dx=A d(xT∗A)/dxT=A d(xT∗A)/dx=AT d(xT∗A∗x)/dx=xT(AT+A) 3. Practice: 4. 矩阵求导计算法则 求导公式(撇号为
# coding: utf-8 # In[2]: from mxnet import nd # In[3]: x = nd.arange(12) x # In[4]: x.shape,x.size # In[5]: x.shape # In[6]: x.size # In[7]: x = x.reshape((3,4)) x # In[9]: x = x.reshape((2,-1)) x # In[11]: x = x.reshape((3,-1)) x # In[12]: nd.zeros(