Return the gradient of `value` with respect to the parameter vector.

For a matrix A, dA^-1/dt = -A^-1 dA/dt A^-1. In this case, we want d(r A^-1 r)/dA = -(A^-1 r)^T (A^-1 r).

The log-determinant of the covariance matrix. This will only be non-``None`` after calling the :func:`compute` method.

An object (following the modeling protocol) that specifies the mean function of the GP. You can safely set this to a scalar, a callab

Sort an N-dimensional list of samples using a KDTree. :param samples: ``(nsamples, ndim)`` The list of samples. This must be a two-d

An array of all parameters (including frozen parameters)

Draw samples from the predictive conditional distribution. You must call :func:`GP.compute` before this function. :param y:

Parameter vector spec.