[llama-mm] Enable kv cache for MultiHeadAttention#6798

larryliu0820 merged 1 commit intomainpytorch/executorch:mainfrom

gh/larryliu0820/55/origpytorch/executorch:gh/larryliu0820/55/origCopy head branch name to clipboard