Enum ModelParadigm 

pub enum ModelParadigm {
    TraditionalMl,
    DeepLearning,
    FineTuning(FineTuneMethod),
    Distillation,
    MoE,
    Ensemble,
}

Model training/inference paradigm

Variants

TraditionalMl

Traditional ML (sklearn-style): logistic regression, random forests, etc.

DeepLearning

Deep learning: neural networks trained from scratch

FineTuning(FineTuneMethod)

Fine-tuning a pretrained model

Distillation

Knowledge distillation (teacher → student)

MoE

Mixture of Experts

Ensemble

Ensemble of multiple models

Implementations

impl ModelParadigm

pub fn lora(rank: u32, alpha: f32) -> Self

Create a LoRA fine-tuning paradigm

pub fn qlora(rank: u32, bits: u8) -> Self

Create a QLoRA fine-tuning paradigm

pub fn typical_memory_multiplier(&self) -> f64

Get typical memory multiplier relative to model size

Returns the factor by which memory usage increases during training compared to inference-only memory.

pub fn typical_training_speedup(&self) -> f64

Get typical training speedup relative to full training

Returns the factor by which training is faster compared to training from scratch.

pub fn typical_quality_retention(&self) -> f64

Get typical quality retention compared to full training

Returns expected quality as a fraction of full fine-tuning quality.

pub fn requires_pretrained(&self) -> bool

Check if this paradigm requires a pretrained model

pub fn is_parameter_efficient(&self) -> bool

Check if this paradigm is parameter-efficient

pub fn batch_size_multiplier(&self) -> f64

Get recommended batch size multiplier

Parameter-efficient methods allow larger batch sizes due to lower memory.

Trait Implementations

impl Clone for ModelParadigm

fn clone(&self) -> ModelParadigm

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ModelParadigm

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

Formats the value using the given formatter.

impl Default for ModelParadigm

fn default() -> ModelParadigm

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for ModelParadigm

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Display for ModelParadigm

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

Formats the value using the given formatter.

impl PartialEq for ModelParadigm

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for ModelParadigm

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 ModelParadigm