Menachem’s paper on systematically accelerating DNN training by using approximate tensor operations got into NeurIPS’21. Congratulations!
