Composite Slice Transformer: An Efficient Transformer with Composition of Multi-Scale Multi-Range Attentions.
Mingu LeeSaurabh PitreTianyu JiangPierre-David LetourneauMatthew J. MorseKanghwan JangJoseph SoriagaParham NoorzadHsin-Pai ChengChristopher LottPublished in: ICLR (2023)