1 paper · 1 filter
Di Cao, Xianchen Wang, Junfeng Zhou +3
Traditional Time Delay Neural Networks (TDNN) have achieved state-of-the-art performance at the cost of high computational complexity and slower inference speed, making them diffic…