Skip to main content

mla_prefill_f32

Function mla_prefill_f32 

Source
pub fn mla_prefill_f32(
    query: &[f32],
    query_pe: &[f32],
    key: &[f32],
    key_pe: &[f32],
    value: &[f32],
    batch: usize,
    query_len: usize,
    key_len: usize,
    heads: usize,
    kv_heads: usize,
    head_dim: usize,
    pe_dim: usize,
    scale: f32,
) -> Vec<f32>