Struct CompressorWeights 

pub struct CompressorWeights {
    pub wkv: DenseWeight,
    pub wgate: DenseWeight,
    pub norm: DenseWeight,
    pub ape: DevicePtr,
    pub ratio: usize,
    pub proj_dim: usize,
    pub is_csa: bool,
    pub pool: DevicePtr,
    pub pool_blocks: usize,
    pub ring: DevicePtr,
    pub prev_win: DevicePtr,
    pub stage: DevicePtr,
}

2026-09-25: DeepSeek-V4 compressed-attention compressor weights, one per layer with a non-zero compress_ratios entry. CSA layers use a 2×ratio overlapping window, HCA layers a single window.

Fields

wkv: DenseWeight

2026-09-25: wkv: [proj_dim, hidden].

wgate: DenseWeight

2026-09-25: wgate: same shape as wkv.

norm: DenseWeight

ape: DevicePtr

2026-09-25: ape, [ratio, proj_dim] FP32.

ratio: usize

2026-09-25: Compression ratio for this layer; is_csa is ratio < 128.

proj_dim: usize

2026-09-25: Output width of wkv/wgate: 2 * head_dim for CSA, head_dim for HCA.

is_csa: bool

2026-09-25: CSA (2×ratio overlapping window) when true, HCA (single window) when false.

pool: DevicePtr

2026-09-25: Persistent flat compressed-KV pool, [pool_blocks × hd_mla] FP8-E4M3 with hd_mla = qk_nope_head_dim + qk_rope_head_dim. Prefill fills blocks [0, n / ratio); decode appends after them.

pool_blocks: usize

2026-09-25: Pool capacity in blocks: max_position_embeddings.div_ceil(ratio).

ring: DevicePtr

2026-09-25: Decode-time ring of compressor inputs, [ratio × hidden] BF16. Each decode token’s normed is written to slot pos % ratio.

prev_win: DevicePtr

2026-09-25: CSA only: the previous completed window’s inputs, [ratio × hidden] BF16. NULL for HCA.

stage: DevicePtr

2026-09-25: CSA only: [2×ratio × hidden] BF16 staging for the overlapped window. NULL for HCA.

Trait Implementations

impl Clone for CompressorWeights

fn clone(&self) -> CompressorWeights

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for CompressorWeights

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

Formats the value using the given formatter.

impl Copy for CompressorWeights