1 paper
Songming Zhang, Yunlong Liang, Shuaibo Wang +4
Knowledge distillation (KD) is a promising technique for model compression in neural machine translation. However, where the knowledge hides in KD is still not clear, which may hin…