pub fn repack_q8_0(data: &[u8]) -> Result<(Vec<u32>, Vec<f32>)>
Layout step for Q8_0: SoA of (i8 words [8/blk], f32 scales) — 36 B / 32 weights = 9.0 bits/weight.
(i8 words [8/blk], f32 scales)