Learning low-precision neural networks without Straight-Through Estimator(STE)

Anonymous