ehartford's picture TheBloke's picture
Change use_cache to True which significantly speeds up inference (#2)
ca45eff