Workshop

ParallelFusion: Towards Maximum Utilization of Mobile GPU for DNN Inference