Enum OverlayTensorKind 

pub enum OverlayTensorKind {
    Base,
    Delta,
    FullSave,
    LoraEmbedA,
    LoraEmbedB,
}

The role an overlay tensor plays in build_overlay.

Variants

Base

token_adapter.base_layer.weight — the adapter’s own [R,h] table.

Delta

token_adapter.trainable_tokens_delta — [T,h] replacement rows.

FullSave

modules_to_save full weight (embed_tokens.weight / lm_head.weight).

LoraEmbedA

Classic low-rank embedding LoRA A factor (Tier-2, load-rejected).

LoraEmbedB

Classic low-rank embedding LoRA B factor (Tier-2, load-rejected).

Trait Implementations

impl Clone for OverlayTensorKind

fn clone(&self) -> OverlayTensorKind

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for OverlayTensorKind

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

Formats the value using the given formatter.

impl PartialEq for OverlayTensorKind

fn eq(&self, other: &OverlayTensorKind) -> 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 Copy for OverlayTensorKind

impl Eq for OverlayTensorKind

impl StructuralPartialEq for OverlayTensorKind