Zobrazeno 1 - 1
of 1
pro vyhledávání: '"Kim, Susang"'
Active research is currently underway to enhance the efficiency of vision transformers (ViTs). Most studies have focused solely on effective token mixers, overlooking the potential relationship with normalization. To boost diverse feature learning, w
Externí odkaz:
http://arxiv.org/abs/2411.18995
