MACaQ: Model-Agnostic Quantization for Efficient Neural Network Inference

MACaQ introduces a fast, model-agnostic quantization method for low-precision weights and activations, reducing costs while preserving accuracy. (Read More)Source