Tag
This paper presents Colla-Q, a bit-allocation framework for quantizing Mixture-of-Experts models using activation entropy to balance expert performance, improving overall model efficiency and reducing calibration dependence.