5 papers
Decode-Branch Transformers: Decoupling the Primary Prefill Path from Additional Decode Computation
Liming Liu, Mingze Wang, Tuo Zhao
As large language models serve ever more requests, cumulative inference cost is growing relative to the one-time cost of training. In typical serving, prompt prefill runs in parall…
BackPlay: Head-Only Look-Back Self-Correction for Diffusion Language Models
Liming Liu, Binxuan Huang, Zixuan Zhang +3
Diffusion Language Models (DLMs) decode multiple tokens in parallel, but aggressive multi-token decoding amplifies cross-token dependency errors and can sharply degrade generation…
COSMOS: A Hybrid Adaptive Optimizer for Memory-Efficient Training of LLMs
Liming Liu, Zhenghao Xu, Zixuan Zhang +5
Large Language Models (LLMs) have demonstrated remarkable success across various domains, yet their optimization remains a significant challenge due to the complex and high-dimensi…
NorMuon: Making Muon more efficient and scalable
Zichong Li, Liming Liu, Chen Liang +2
The choice of optimizer significantly impacts the training efficiency and computational costs of large language models (LLMs). Recently, the Muon optimizer has demonstrated promisi…
A Minimalist Example of Edge-of-Stability and Progressive Sharpening
Liming Liu, Zixuan Zhang, Simon Du +1
Recent advances in deep learning optimization have unveiled two intriguing phenomena under large learning rates: Edge of Stability (EoS) and Progressive Sharpening (PS), challengin…