I would like to ask does anyone have idea how to provide two sequences within a transformer model( multi head attention to make it work like cross attention), I tried many times but doesnt understa...