DirectSolver

class tinygp.solvers.DirectSolver(kernel: kernels.Kernel, X: JAXArray, noise: Noise, *, covariance: Any | None = None)

Bases: Solver

A direct solver that uses jax’s built in Cholesky factorization

You generally won’t instantiate this object directly but, if you do, you’ll probably want to use the DirectSolver.init() method instead of the usual constructor.

condition(kernel: Kernel, X_test: TypeAliasForwardRef('tinygp.helpers.JAXArray') | None, noise: Noise) → Any

Compute the covariance matrix for a conditional GP

Parameters:

• kernel – The kernel for the covariance between the observed and predicted data.

• X_test – The coordinates of the predicted points. Defaults to the input coordinates.

• noise – The noise model for the predicted process.

covariance() → tinygp.helpers.JAXArray

The evaluated covariance matrix

dot_triangular(y: tinygp.helpers.JAXArray) → tinygp.helpers.JAXArray

Compute a matrix product with the lower triangular linear system

If the covariance matrix is K = L @ L.T for some lower triangular matrix L, this method returns L @ y for some y.

normalization() → tinygp.helpers.JAXArray

The multivariate normal normalization constant

This should be (log_det + n*log(2*pi))/2, where n is the size of the covariance matrix, and log_det is the log determinant of the matrix.

solve_triangular(y: tinygp.helpers.JAXArray, *, transpose: bool = False) → tinygp.helpers.JAXArray

Solve the lower triangular linear system defined by this solver

If the covariance matrix is K = L @ L.T for some lower triangular matrix L, this method solves L @ x = y for some y. If the transpose parameter is True, this instead solves L.T @ x = y.

variance() → tinygp.helpers.JAXArray

The diagonal of the covariance matrix