VecQ: Minimal Loss DNN Model Compression With Vectorized Weight Quantization | IEEE Journals & Magazine | IEEE Xplore