Reduce allocation overhead in quantized sdpa#16119

kimishpatel merged 1 commit intomainpytorch/executorch:mainfrom

gh/kimishpatel/202/origpytorch/executorch:gh/kimishpatel/202/origCopy head branch name to clipboard