Year: 2025
Type: conference-paper
Abstract: Mixture-of-Experts (MoE) shines brightly in large language models (LLMs) and demonstrates outstanding performance in plentiful natural language processing tasks.However, existing methods transforming LLMs from dense to MoE face significant data requirements and typically rely on large-scale post-training.In this paper, we propose Upcycling Instruction Tuning (UpIT), a data-efficient approach for tuning a dense pre-trained model into a MoE instruction model.Specifically, we first point out that intermediate checkpoints during instruction tuning of the dense model are naturally suitable for specialized experts, and then propose an expert expansion stage to flexibly achieve models with flexible numbers of experts, where genetic algorithm and parameter merging are introduced to ensure sufficient diversity of new extended experts.To ensure that each specialized expert in the MoE model works as expected, we select a small amount of seed data that each expert excels to preoptimize the router.Extensive experiments with various data scales and upcycling settings demonstrate the outstanding performance and data efficiency of UpIT, as well as stable improvement in expert or data scaling.Further analysis reveals the importance of ensuring expert diversity in upcycling.
Authors: Tingfeng Hui, Zhenyu Zhang, Shuohuan Wang, Yu Sun, Hua Wu, Sen Su
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Awards: 隐私保护的数据集成与共享关键技术研究, G2993257473
OpenAlex ID: W4412945365