Struct LoraMapper 

pub struct LoraMapper { /* private fields */ }

A module mapper that attaches LoRA adapters to 2-D weight parameters.

Apply it the same way as quantization:

let model = model.map(&mut LoraMapper::new(LoraConfig::new(8, 16.0)));

Each rank-2 float weight is frozen and given a trainable low-rank adapter; other parameters are left untouched. No model or layer code needs to change — the same Linear (and any other module) keeps working, now producing base + scale * (a @ b) for adapted weights.

Implementations

impl LoraMapper

pub fn new(config: LoraConfig) -> Self

Create a new mapper from the given configuration.

pub fn for_group(self, group: ParamGroup) -> Self

Specify a parameter group on which to apply the LoRA.

Trait Implementations

impl Clone for LoraMapper

fn clone(&self) -> LoraMapper

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LoraMapper

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

Formats the value using the given formatter.

impl ModuleMapper for LoraMapper

fn enter_module(&mut self, name: &str, _container_type: &str)

Called when entering a submodule.

fn exit_module(&mut self, _name: &str, _container_type: &str)

Called when exiting a submodule.

fn map_float<const D: usize>( &mut self, param: Param<Tensor<D>>, ) -> Param<Tensor<D>>

Map a float parameter in the module.

fn map_int<const D: usize>( &mut self, param: Param<Tensor<D, Int>>, ) -> Param<Tensor<D, Int>>

Map an int parameter in the module.

fn map_bool<const D: usize>( &mut self, param: Param<Tensor<D, Bool>>, ) -> Param<Tensor<D, Bool>>

Map a bool parameter in the module.