def algo_jaco(*args, **kwargs):
var = UTPM.init_hessian(args[narg])
diff_args = list(args) # We are making a copy of args
diff_args[narg] = var
diff_args[-1]= var
diff_args = tuple(diff_args)
return UTPM.extract_hessian(rhfenergy(*(diff_args)))
def gradient(f, theta, x):
a = UTPM.init_hessian(x)
y = f(theta, a)
res = UTPM.extract_hessian(2, y) #UTPM.extract_jacobian(y)
out = res.diagonal(1)
return out[0]
def _forward(self, x, *args, **kwds):
x = np.atleast_1d(x)
tmp = UTPM.init_hessian(x)
y = self.fun(tmp, *args, **kwds)
return UTPM.extract_hessian(len(x), y)
