Return to Article Details Characterizing Hardware Utilization on Edge Devices when Inferring Compressed Deep Learning Models Download Download PDF