llama.cpp

Using `uint8x16_t` as the return type for `ggml_vqtbl1q_u8`

#5894

Merged