Permute, Quantize, and Fine-tune: Efficient Compression of Neural Networks | IEEE Conference Publication | IEEE Xplore