rope_mrope_interleaved_k_only

Function rope_mrope_interleaved_k_only 

Source
pub fn rope_mrope_interleaved_k_only(
    gpu: &dyn GpuBackend,
    kernel: KernelHandle,
    k: DevicePtr,
    pos_t: DevicePtr,
    pos_h: DevicePtr,
    pos_w: DevicePtr,
    seq_len: u32,
    num_kv_heads: u32,
    head_dim: u32,
    rotary_dim: u32,
    theta: f32,
    stream: u64,
) -> Result<()>
Expand description

MRoPE for K only. Used when Q was already rotated by a fused Q prefill kernel.