Yes we have entire pipeline (detection -> embedding -> clustering) for faces as well as ability to extract text tags using models such as Open Images.