Jiaxing Shi

Hbp: Hierarchically balanced pruning and accelerator co-design for efficient dnn inference