Gate weights cache on runtime option instead of compile-time macro (#19603)#19603

GregoryComer merged 2 commits into

pytorch:mainpytorch/executorch:mainfrom

hboyraz:export-D105123995hboyraz/executorch:export-D105123995Copy head branch name to clipboard