SAVE: Sparsity-Aware Vector Engine for Accelerating DNN Training and Inference on CPUs | IEEE Conference Publication | IEEE Xplore