Mixture-of-experts (MoE) transformers, the standard backbone for modern large-scale deep learning, employ conditional computation: only a sparse subset of mo…
机构:港大
来源:arXiv 2610.31818 | AI4Papers 论文推荐平台