INQ(incremental network quantization：towards lossless CNNs with low-precision weights

incremental network quantization：towards lossless CNNs with low-precision weights 实验前提：a.一个预训练的模型，模型已收敛                 b.训练这个模型训练集，验证集，测试集 大体三个步骤：weights partition、group-wise quantization 、re-trainin

>>阅读原文<<