https://arxiv.org/abs/2103.10619
Scalable Visual Transformers with Hierarchical Pooling (Zizheng Pan, Bohan Zhuang, Jing Liu, Haoyu He, Jianfei Cai)
funnel transformer나 pyramid vision transformer와 비슷하게 pooling을 사용해서 vision transformer를 효율화. 이쪽은 1d pooling으로 해결.
#vision_transformer