cache-type-k/v not support Q4_K_4 and Q8R16

[ed-hch]

Is it possible/beneficial to add support on the Ampere's format for cache as well?
-ctk, --cache-type-k TYPE KV cache data type for K
allowed values: f32, f16, bf16, q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1
(default: f16)
(env: LLAMA_ARG_CACHE_TYPE_K)
to show complete usage, run with -h