8 papers
ReCoVer: Resilient LLM Pre-Training System via Fault-Tolerant Collective and Versatile Workload
Ziyue Liu, Zhengyang Wang, Ruijie Zhang +7
Pre-training large language models on massive GPU clusters has made hardware faults routine rather than rare, driving the need for resilient training systems. Yet existing framewor…
Muon: Boosting Muon via Adaptive Second-Moment Preconditioning
Ziyue Liu, Ruijie Zhang, Zhengyang Wang +4
Muon has emerged as a promising optimizer for large-scale foundation model pre-training by exploiting the matrix structure of neural network updates through iterative orthogonaliza…
TEON: Tensorized Orthonormalization Beyond Layer-Wise Muon for Large Language Model Pre-Training
Ruijie Zhang, Yequan Zhao, Ziyue Liu +5
The Muon optimizer has demonstrated strong empirical performance in pre-training large language models by performing matrix-level gradient (or momentum) orthogonalization in each l…
MUON+: Towards More Effective Muon via One Additional Normalization Step for LLM Pre-training
Ruijie Zhang, Yequan Zhao, Ziyue Liu +4
Muon has recently emerged as a strong optimizer for large language model pre-training, orthogonalizing the momentum matrix via Newton--Schulz polar iterations. A natural intuition…
BOOST: BOttleneck-Optimized Scalable Training Framework for Low-Rank Large Language Models
Zhengyang Wang, Ziyue Liu, Ruijie Zhang +5
The scale of transformer model pre-training is constrained by the increasing computation and communication cost. Low-rank bottleneck architectures offer a promising solution to sig…
SkipKV: Selective Skipping of KV Generation and Storage for Efficient Inference with Large Reasoning Models
Jiayi Tian, Seyedarmin Azizi, Yequan Zhao +7
Large reasoning models (LRMs) often incur significant key-value (KV) cache overhead, due to their linear growth with the verbose chain-of-thought (CoT) reasoning. This incurs both…
LaX: Boosting Low-Rank Training of Foundation Models via Latent Crossing
Ruijie Zhang, Ziyue Liu, Zhengyang Wang +1
Training foundation models such as ViTs and LLMs requires tremendous computing cost. Low-rank matrix or tensor factorization offers a parameter-efficient alternative, but often dow…
CoLA: Compute-Efficient Pre-Training of LLMs via Low-Rank Activation
Ziyue Liu, Ruijie Zhang, Zhengyang Wang +7
The full-size MLPs and the projection layers in attention introduce tremendous model sizes of large language models (LLMs), consuming extensive computational resources in pre-train…