Struct HyperparameterSpace 

pub struct HyperparameterSpace {
    pub learning_rate: (f32, f32),
    pub beta1: (f32, f32),
    pub beta2: (f32, f32),
    pub weight_decay: (f32, f32),
    pub epsilon: (f32, f32),
    pub batch_sizes: Vec<usize>,
    pub log_scale_lr: bool,
    pub custom_params: HashMap<String, (f32, f32)>,
}

Hyperparameter search space definition

Fields

learning_rate: (f32, f32)

Learning rate bounds (min, max)

beta1: (f32, f32)

Beta1 momentum bounds

beta2: (f32, f32)

Beta2 momentum bounds

weight_decay: (f32, f32)

Weight decay bounds

epsilon: (f32, f32)

Epsilon bounds

batch_sizes: Vec<usize>

Batch size options (discrete)

log_scale_lr: bool

Whether to use logarithmic scaling for learning rate

custom_params: HashMap<String, (f32, f32)>

Custom parameter ranges for specific optimizers

Implementations

impl HyperparameterSpace

pub fn for_transformers() -> Self

Create search space optimized for transformer models

pub fn for_vision() -> Self

Create search space for vision models

pub fn for_scientific_computing() -> Self

Create search space for scientific computing

Trait Implementations

impl Clone for HyperparameterSpace

fn clone(&self) -> HyperparameterSpace

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for HyperparameterSpace

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

Formats the value using the given formatter.

impl Default for HyperparameterSpace

fn default() -> Self

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for HyperparameterSpace

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for HyperparameterSpace

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

where __S: Serializer,

Serialize this value into the given Serde serializer.