Evaluate and convergence options (shared by evaluate + benchmark). More...

#include <EvaluateParams.hpp>

Classes
struct	Convergence
	Convergence control. More...

Public Member Functions
	EvaluateParams ()=default

	EvaluateParams (nlohmann::json const &config)
	Construct from a JSON config object.

bool	convergence_enabled () const
	Whether adaptive convergence stopping is active (disabled when iterations is set).

nlohmann::json	overrides () const
	JSON with only the fields explicitly set (for scenario overrides).

void	resolve_defaults ()
	Fill unset optional fields with their default values.

nlohmann::json	to_json () const
	Serialize to JSON config.

Static Public Member Functions
static void	validate (nlohmann::json const &config, std::vector< std::string > &errors)
	Validate evaluate-related fields in a JSON config.

Public Attributes
struct ppforest2::cli::EvaluateParams::Convergence	convergence

std::string	export_path

bool	fixed_seed = false
	Use the user-provided seed unchanged for every iteration.

std::optional< int >	iterations
	Fixed count (disables convergence when set > 0).

std::optional< float >	train_ratio

int	warmup = 0
	Warmup iterations discarded before measuring.

Detailed Description

Evaluate and convergence options (shared by evaluate + benchmark).

Constructor & Destructor Documentation

◆ EvaluateParams() [1/2]

ppforest2::cli::EvaluateParams::EvaluateParams ( )

default

◆ EvaluateParams() [2/2]

ppforest2::cli::EvaluateParams::EvaluateParams ( nlohmann::json const & config )

explicit

Construct from a JSON config object.

Member Function Documentation

◆ convergence_enabled()

bool ppforest2::cli::EvaluateParams::convergence_enabled ( ) const

inline

Whether adaptive convergence stopping is active (disabled when iterations is set).

◆ overrides()

nlohmann::json ppforest2::cli::EvaluateParams::overrides ( ) const

JSON with only the fields explicitly set (for scenario overrides).

◆ resolve_defaults()

void ppforest2::cli::EvaluateParams::resolve_defaults ( )

Fill unset optional fields with their default values.

◆ to_json()

nlohmann::json ppforest2::cli::EvaluateParams::to_json ( ) const

Serialize to JSON config.

◆ validate()

static void ppforest2::cli::EvaluateParams::validate	(	nlohmann::json const &	config,
		std::vector< std::string > &	errors )

static

Validate evaluate-related fields in a JSON config.

Member Data Documentation

◆ convergence

struct ppforest2::cli::EvaluateParams::Convergence ppforest2::cli::EvaluateParams::convergence

◆ export_path

std::string ppforest2::cli::EvaluateParams::export_path

◆ fixed_seed

bool ppforest2::cli::EvaluateParams::fixed_seed = false

Use the user-provided seed unchanged for every iteration.

Default (false): each iteration draws a fresh seed from rng so models genuinely differ across iterations. std_*_error reflects real algorithmic variance; std_time_ms mixes system noise with model-shape variance.

true: every iteration trains the same model. std_*_error is tautologically zero; std_time_ms reflects pure timing noise (cache, scheduling, allocator jitter). Useful when benchmarking the timing stability of a single configuration.

◆ iterations

std::optional<int> ppforest2::cli::EvaluateParams::iterations

Fixed count (disables convergence when set > 0).

◆ train_ratio

std::optional<float> ppforest2::cli::EvaluateParams::train_ratio

◆ warmup

int ppforest2::cli::EvaluateParams::warmup = 0

Warmup iterations discarded before measuring.

The documentation for this struct was generated from the following file:

core/src/cli/EvaluateParams.hpp

Classes

Public Member Functions

Static Public Member Functions

Public Attributes