Struct ZeroShotConfig 

pub struct ZeroShotConfig {

    pub num_probe_points: usize,
    pub probe_step: f64,
    pub num_curvature_directions: usize,
    pub num_random_curvature_dirs: usize,
    pub finite_diff_eps: f64,
    pub sparsity_rel_threshold: f64,
    pub noise_repeats: usize,
    pub classifier_epochs: usize,
    pub classifier_lr: f64,
    pub classifier_l2: f64,
    pub regressor_epochs: usize,
    pub regressor_lr: f64,
    pub regressor_l2: f64,
    pub min_learning_rate: f64,
    pub max_learning_rate: f64,
    pub seed: u64,
}

Configuration for meta-feature extraction and meta-model training.

Fields

num_probe_points: usize

Number of points visited along the normalized-descent probe path (>= 1).

probe_step: f64

Step size of the normalized-descent probe walk.

num_curvature_directions: usize

Maximum number of coordinate directions used for curvature probing.

num_random_curvature_dirs: usize

Number of additional random directions used for curvature probing.

finite_diff_eps: f64

Finite-difference step (scaled by the parameter RMS) for curvature estimation.

sparsity_rel_threshold: f64

Relative threshold below which a gradient component counts as “zero” for sparsity.

noise_repeats: usize

Number of repeated gradient samples used to estimate gradient noise (>= 0).

classifier_epochs: usize

Number of full-batch gradient-descent epochs for the classifier.

classifier_lr: f64

Classifier learning rate.

classifier_l2: f64

Classifier L2 regularization strength.

regressor_epochs: usize

Number of full-batch gradient-descent epochs for the learning-rate regressor.

regressor_lr: f64

Regressor learning rate.

regressor_l2: f64

Regressor L2 regularization strength.

min_learning_rate: f64

Minimum learning rate the regressor may recommend.

max_learning_rate: f64

Maximum learning rate the regressor may recommend.

seed: u64

Seed for the random curvature directions (keeps extraction deterministic).

Implementations

impl ZeroShotConfig

pub fn validate(&self) -> Result<()>

Validate the configuration, returning an error describing the first problem found.

Trait Implementations

impl Clone for ZeroShotConfig

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ZeroShotConfig

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for ZeroShotConfig

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for ZeroShotConfig

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for ZeroShotConfig

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for ZeroShotConfig

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for ZeroShotConfig