Follow
DE$^3$-BERT: Distance-Enhanced Early Exiting for BERT based on Prototypical Networks https://arxiv.org/abs/2402.05948 #cs.LG #cs.CL
DE$^3$-BERT: Distance-Enhanced Early Exiting for BERT based on Prototypical Networks https://arxiv.org/abs/2402.05948 #cs.LG #cs.CL