Skip to main content

encode_row_q8_0

Function encode_row_q8_0 

Source
pub fn encode_row_q8_0(src: &[f32], out: &mut Vec<u8>) -> Option<()>
Expand description

Encodes a whole row (or any slice whose length is a multiple of Q8_0_BLOCK_ELEMS) into Q8_0 blocks, appending to out.

Returns None when src.len() is not a multiple of the block size. llama.cpp asserts the same condition and its Q8_0 path has no fallback type, so a row that cannot be tiled is a refusal, never a zero-padded block: padding changes the row length the reader computes from the shape, and the file would decode shifted.