I am trying to compute the partial derivative of the log pdf of a normal distribution with autograd as in the following code :
import autograd.scipy.stats as stat
import autograd.numpy as np
from autograd import grad
def f(x, lc, sc):
return stat.norm.logpdf(x, loc= lc, scale = sc)
grad_f = grad(f, argnum=1)
grad_f(-1.0, 1.)
and I get the following error message :
"ValueError: setting an array element with a sequence".
I believe that autograd knows how to handle gradients of the norm.logpdf since it is an in-built function and no other vector assignments have been made...