Direct Quantization for Training Highly Accurate Low Bit-width Deep Neural Networks

MQ
Re-assign community