Struct LoRAAdapter 

pub struct LoRAAdapter {
    pub context_type: ContextType,
    pub meta: LoRAMetadata,
    pub lora_a: LoRATensor,
    pub lora_b: LoRATensor,
}

A loaded LoRA adapter ready for CPU or GPU application.

Fields

context_type: ContextType

meta: LoRAMetadata

lora_a: LoRATensor

Down-projection: [rank × n_in]

lora_b: LoRATensor

Up-projection: [n_out × rank]

Implementations

impl LoRAAdapter

pub fn apply_cpu( &self, input: &[f32], output: &mut [f32], ) -> Result<(), LoRAError>

Apply the LoRA delta additively to output.

Implements output += lora_b @ (lora_a @ input) * scaling.

• input must have length meta.n_in
• output must have length meta.n_out (values are preserved and incremented)

pub fn compute_delta(&self, input: &[f32]) -> Result<Vec<f32>, LoRAError>

Compute the LoRA delta vector without applying it. Returns delta of length meta.n_out.

Trait Implementations

impl Clone for LoRAAdapter

fn clone(&self) -> LoRAAdapter

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LoRAAdapter

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

Formats the value using the given formatter.