Enum LayerType 

pub enum LayerType {
    FullAttention,
    SlidingAttention,
    LinearAttention,
    Moe,
    SparseAttention,
}

Variants

FullAttention

SlidingAttention

LinearAttention

Moe

2026-09-26: A layer with no mixer, only the routed-expert FFN (Nemotron-H pattern E, Puzzle moe).

SparseAttention

2026-09-26: Sparse attention (deepseek_sparse_attention): a full-rank mixer whose visible key set an indexer selects per query at run time.

Implementations

impl LayerType

pub fn is_attention(self) -> bool

pub fn hf_name(self) -> &'static str

Trait Implementations

impl Clone for LayerType

fn clone(&self) -> LayerType

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LayerType

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for LayerType

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for LayerType

fn eq(&self, other: &LayerType) -> 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 LayerType

impl Eq for LayerType

impl StructuralPartialEq for LayerType