intelligence.zenith_tune.strategies.grid

Grid search strategy for hyperparameter optimization.

GridStrategy Objects

@StrategyRegistry.register("grid")
class GridStrategy(TuningStrategy)

Grid search strategy.

This strategy exhaustively searches all combinations of parameters in the search space. For continuous parameters (Int/Float), the range is discretized based on step size or a default grid resolution.

Naturally terminates when all grid points have been evaluated.

Example:

strategy = GridStrategy() strategy.optimize(search_space, direction, eval_fn)

__init__

def __init__(grid_points: int = 5) -> None

Initialize the grid strategy.

Arguments:

• grid_points - Number of grid points for parameters without explicit step size.