Struct LoraAdapter 

pub struct LoraAdapter {
    pub a: Param<Tensor<2>>,
    pub b: Param<Tensor<2>>,
    pub scale: f64,
}

A LoRA (Low-Rank Adaptation) adapter attached to a frozen weight parameter.

When present on a Param<Tensor<2>>, the parameter materializes its effective value as base + scale * (a @ b), where base is the frozen (and optionally quantized) weight and a/b are the trainable low-rank factors. The frozen base is the stored value of the parameter; the adapter factors are surfaced to the optimizer, autodiff and record systems as regular parameters with their own ParamIds through the module visitor/mapper traversal.

Fields

a: Param<Tensor<2>>

Down-projection factor with shape [d_in, rank] (trainable).

b: Param<Tensor<2>>

Up-projection factor with shape [rank, d_out] (trainable).

scale: f64

Scaling factor applied to the low-rank product, typically alpha / rank.

Implementations

impl LoraAdapter

pub fn delta(&self) -> Tensor<2>

Compute the low-rank delta scale * (a @ b) with shape [d_in, d_out].

a and b are read through Param::val, so the delta always reflects the current (optimizer-updated) factors and keeps them as autodiff leaves for backpropagation.

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<(), Error>

Formats the value using the given formatter.

impl Display for LoraAdapter

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

Formats the value using the given formatter.

impl Module for LoraAdapter

fn num_params(&self) -> usize

Get the number of parameters the module has, including all of its sub-modules, without initializing the ones not initialized yet.

fn visit<Visitor>(&self, visitor: &mut Visitor)

where Visitor: ModuleVisitor,

Visit each parameter and module-owned control value with a visitor.

fn map<Mapper>(self, mapper: &mut Mapper) -> LoraAdapter

where Mapper: ModuleMapper,

Map each parameter and module-owned control value with a mapper.

fn collect_devices(&self, devices: Vec<Device>) -> Vec<Device>

Return all the devices found in the underneath module tree added to the given vector without duplicates.

fn to_device(self, device: &Device) -> LoraAdapter

Move the module and all of its sub-modules to the given device.

fn fork(self, device: &Device) -> LoraAdapter

Fork the module and all of its sub-modules to the given device.

fn valid(&self) -> LoraAdapter

Returns a validation snapshot without autodiff or active training flags.

fn train(self) -> LoraAdapter

Enables autodiff and restores configured training state throughout the module.

fn materialize(self) -> LoraAdapter

Fold parameter reparameterizations into their effective values and remove them.

fn devices(&self) -> Vec<Device>

Return all the devices found in the underneath module tree without duplicates.

fn set_require_grad(self, require_grad: bool) -> Self

Set whether every floating-point tensor parameter in the module tree requires gradients.

fn set_require_grad_group(self, group: ParamGroup, require_grad: bool) -> Self

Set whether matched floating-point tensor parameters require gradients.

fn no_grad(self) -> Self

Disable gradients for every floating-point tensor parameter in the module tree.

fn freeze(self) -> Self

Freeze the whole module tree.

fn unfreeze(self) -> Self

Unfreeze the whole module tree.

fn freeze_group(self, group: ParamGroup) -> Self

Freeze every module-owned value in the given group, leaving the rest of the module untouched.

fn unfreeze_group(self, group: ParamGroup) -> Self

Unfreeze every module-owned value in the given group, leaving the rest of the module untouched.

fn quantize_weights(self, quantizer: &mut Quantizer) -> Self

Quantize the weights of the module.

fn quantize_weights_group( self, quantizer: &mut Quantizer, group: ParamGroup, ) -> Self

Quantize the weights of the given parameter group.

fn apply_reparameterization<R>(self, reparameterizer: R) -> Self

where Self: Sized, R: Reparameterizer,

Attach reparameterizations using the given Reparameterizer.

fn apply_lora(self, lora: Lora) -> Self

where Self: Sized,

Attach LoRA adapters to the module’s 2-D weights, freezing all base tensor parameters.

fn apply_qlora(self, qlora: QLora) -> Self

where Self: Sized,

Apply QLoRA to the module: quantize the (frozen) base tensor parameters and attach trainable LoRA adapters to 2-D weights.

fn into_record(self) -> ModuleRecord

where Self: Sized,

Collect this module’s tensor parameters into a ModuleRecord.

fn into_record_group(self, group: ParamGroup) -> ModuleRecord

where Self: Sized,

Collect the tensor parameters group names into a ModuleRecord.

fn try_load_record(self, record: ModuleRecord) -> Result<Self, RecordError>

where Self: Sized,

Apply a ModuleRecord to this module, returning the loaded module.

fn load_record(self, record: ModuleRecord) -> Self

where Self: Sized,

Apply a ModuleRecord to this module, consuming and returning it.

fn save_file<P>(self, path: P) -> Result<(), RecordError>

where P: AsRef<Path>, Self: Sized,

Save this module’s parameters to a burnpack file on disk.

fn load_file<P>(self, path: P) -> Self

where P: AsRef<Path>, Self: Sized,

Load this module’s parameters from a burnpack file on disk, returning the loaded module.

fn try_load_file<P>(self, path: P) -> Result<Self, RecordError>

where P: AsRef<Path>, Self: Sized,

Fallible variant of load_file.

impl ModuleDisplay for LoraAdapter

fn format(&self, passed_settings: DisplaySettings) -> String

Formats the module with provided display settings.

fn custom_settings(&self) -> Option<DisplaySettings>

Custom display settings for the module.

fn custom_content(&self, _content: Content) -> Option<Content>

Custom attributes for the module.

impl ModuleDisplayDefault for LoraAdapter

fn content(&self, content: Content) -> Option<Content>

Attributes of the module used for display purposes.

fn num_params(&self) -> usize

Gets the number of the parameters of the module.

impl Reparameterization for LoraAdapter

const NAME: &'static str = "lora"

Stable path component used for the reparameterization’s nested parameters.

fn apply<const D: usize>(&self, base: Tensor<D>) -> Tensor<D>

Apply the transformation to the stored base, returning the effective parameter value.