[TRTLLM-5530][BREAKING CHANGE] refactor: unify KvCacheConfig in LLM class for pytorch backend#5752

Superjomn merged 3 commits intoNVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

Superjomn:change-kvcache-configSuperjomn/TensorRT-LLM:change-kvcache-configCopy head branch name to clipboard