Hardware-Aware DNN Compression via Diverse Pruning and Mixed-Precision Quantization | IEEE Journals & Magazine | IEEE Xplore