Quantizing deep convolutional networks for efficient inference: A whitepaper

Raghuraman Krishnamoorthi

We present an overview of techniques for quantizing convolutional neural networks for inference with integer weights and activations. Per-channel quantization of weights and per-la…