Search Results for author:
Found 2 papers, 1 papers with code
MoDE: A Mixture-of-Experts Model with Mutual Distillation among the Experts
However, the gate's routing mechanism also gives rise to narrow vision: the individual MoE's expert fails to use more samples in learning the allocated sub-task, which in turn limits the MoE to further improve its generalization ability.
Lookahead: An Inference Acceleration Framework for Large Language Model with Lossless Generation Accuracy
Hence, this paper presents a generic framework for accelerating the inference process, resulting in a substantial increase in speed and cost reduction for our RAG system, with lossless generation accuracy.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
