attn_fp8_twin_bytes

Function attn_fp8_twin_bytes 

Source
pub fn attn_fp8_twin_bytes(
    want: Fp8TwinSet,
    q_n: usize,
    kv_n: usize,
    o_k: usize,
    hidden: usize,
) -> usize
Expand description

The FP8 prefill twins want selects, for one full-attention layer.