prepare_optimization

Optimization.prepare_optimization(initial_params: ndarray | None = None) → ndarray

Validate inputs and reset per-run state for an optimization run.

Builds nothing optimizer-specific — the internal iteration callback is constructed inside _run_optimization() so it can close over the caching wrappers without any cross-method state.

Parameters:

initial_paramsnp.ndarray, optional

Initial parameter values. If None, uses default initialization from the parametrization (default: None).

Returns:

np.ndarray

Validated initial parameters.

Raises:

ValueError

If initial_params is provided but has the wrong shape.