A Survey of Quantization Methods for Efficient Neural Network Inference