pub fn rotate_blocks_partial<const D: usize, const DB: usize>(
v: Tensor<D>,
q: Tensor<DB>,
rope_width: usize,
) -> Tensor<D>Expand description
Apply a per-block quaternion rotation to the first rope_width entries of
the state_rank axis (a multiple of 4); the remainder passes through. The
quaternion analogue of apply_rope_partial.
q has one quaternion per rotated block (rope_width / 4 of them). DB
must equal D + 1.