7 papers
SpiralFormer: Looped Transformers Can Learn Hierarchical Dependencies via Multi-Resolution Recursion
Chengting Yu, Xiaobo Shu, Yadao Wang +8
Recursive (looped) Transformers decouple computational depth from parameter depth by repeatedly applying shared layers, providing an explicit architectural primitive for iterative…
MeSH: Memory-as-State-Highways for Recursive Transformers
Chengting Yu, Xiaobo Shu, Yadao Wang +8
Recursive transformers reuse parameters and iterate over hidden states multiple times, decoupling compute depth from parameter depth. However, under matched compute, recursive mode…
Expert Divergence Learning for MoE-based Language Models
Jiaang Li, Haibin Chen, Langming Liu +9
The Mixture-of-Experts (MoE) architecture is a powerful technique for scaling language models, yet it often suffers from expert homogenization, where experts learn redundant functi…
Reinforcement Learning Optimization for Large-Scale Learning: An Efficient and User-Friendly Scaling Library
Weixun Wang, Shaopan Xiong, Gengru Chen +38
We introduce ROLL, an efficient, scalable, and user-friendly library designed for Reinforcement Learning Optimization for Large-scale Learning. ROLL caters to three primary user gr…
Weight Spectra Induced Efficient Model Adaptation
Chongjie Si, Xuankun Yang, Muqing Liu +5
Large-scale foundation models have demonstrated remarkable versatility across a wide range of downstream tasks. However, fully fine-tuning these models incurs prohibitive computati…
MAP: Revisiting Weight Decomposition for Low-Rank Adaptation
Chongjie Si, Zhiyi Shi, Yadao Wang +3
The rapid development of large language models has revolutionized natural language processing, but their fine-tuning remains computationally expensive, hindering broad deployment.…