Per-channel Quantization Level Allocation for Quantizing Convolutional Neural Networks | IEEE Conference Publication | IEEE Xplore