8 papers
SLAI T-Rex: Full-Parameter Post-training of the DeepSeek-V4 Family on Ascend SuperPOD
Dongfang Li, Xiaodong Luo, Ruoyu Sun +64
Full-parameter post-training of trillion-parameter-scale MoE models introduces substantial system-level challenges for large-scale distributed training, including severe memory pre…
Adam Converges Without Any Modification On Update Rules
Yushun Zhang, Bingran Li, Congliang Chen +2
Adam is the default algorithm for training neural networks, including large language models (LLMs). However, \citet{reddi2019convergence} provided an example that Adam diverges, ra…
MoFO: Momentum-Filtered Optimizer for Mitigating Forgetting in LLM Fine-Tuning
Yupeng Chen, Senmiao Wang, Yushun Zhang +5
Large language models (LLMs) have demonstrated remarkable capabilities across a wide range of tasks. Typically, LLMs are first pre-trained on large corpora and subsequently fine-tu…
Knapsack RL: Unlocking Exploration of LLMs via Optimizing Budget Allocation
Ziniu Li, Congliang Chen, Tianyun Yang +5
Large Language Models (LLMs) can self-improve through reinforcement learning, where they generate trajectories to explore and discover better solutions. However, this exploration p…
Towards Quantifying the Hessian Structure of Neural Networks
Zhaorui Dong, Yushun Zhang, Jianfeng Yao +1
Empirical studies reported that the Hessian matrix of neural networks (NNs) exhibits a near-block-diagonal structure, yet its theoretical foundation remains unclear. In this work,…
Rethinking Data Mixture for Large Language Models: A Comprehensive Survey and New Perspectives
Yajiao Liu, Congliang Chen, Junchi Yang +1
Training large language models with data collected from various domains can improve their performance on downstream tasks. However, given a fixed training budget, the sampling prop…