1 paper · 1 filter
Zhimin Tang, Linkai Luo, Bike Xie +4
Since sparse neural networks usually contain many zero weights, these unnecessary network connections can potentially be eliminated without degrading network performance. Therefore…