[Fix] Remove unused code to reduce binary size (#181)

* clean-up

* fix lint

* fix lint

src/turbomind/models/llama/prefix_cache.h

-// Copyright (c) OpenMMLab. All rights reserved.
-
-#include <cuda_fp16.h>
-
-template<typename T>
-void invokeInsertKeyCache(T* key_cache, const T* src, int L, int H, int Dx, int s, int X, int S, cudaStream_t st);
-
-template<typename T>
-void invokeInsertValueCache(T* value_cache, const T* src, int L, int H, int s, int D, int S, cudaStream_t st);