pub struct AttnV41LayerWeights {
pub role: LayerRole,
pub sink: DevicePtr,
pub wq_a: AttnMat,
pub q_norm: DevicePtr,
pub wq_b: AttnMat,
pub wkv: AttnMat,
pub kv_norm: DevicePtr,
pub wo_a: AttnMat,
pub wo_b: AttnMat,
pub comp: Option<CompressorWeightsGpu>,
pub idx: Option<IndexerWeightsGpu>,
}Expand description
2026-09-25: One layer’s resident attention weights. Projections are
[out, in]: wo_a is [groups * o_rank, gw], wo_b [dim, groups * o_rank]. sink ([n_heads]), q_norm ([q_rank]) and kv_norm
([hd]) are f32.
Fields§
§role: LayerRole§sink: DevicePtr§wq_a: AttnMat§q_norm: DevicePtr§wq_b: AttnMat§wkv: AttnMat§kv_norm: DevicePtr§wo_a: AttnMat§wo_b: AttnMat§comp: Option<CompressorWeightsGpu>§idx: Option<IndexerWeightsGpu>Auto Trait Implementations§
impl Freeze for AttnV41LayerWeights
impl RefUnwindSafe for AttnV41LayerWeights
impl Send for AttnV41LayerWeights
impl Sync for AttnV41LayerWeights
impl Unpin for AttnV41LayerWeights
impl UnwindSafe for AttnV41LayerWeights
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more