12 papers
Erase-then-Delta Attention: Decoupling Erase and Write Addresses in Delta-Rule Linear Attention
Xiao Li, Chengruidong Zhang, Hao Luo +15
Delta-rule linear attention improves recurrent memory updates by correcting what is already stored at the current write address before writing new content. However, the active corr…
Breaking Entropy Bounds: Accelerating RL Training via MTP with Rejection Sampling
Yucheng Li, Huiqiang Jiang, Yang Xu +14
Reinforcement learning (RL) has become a key component in modern large language models, yet the rollout stage remains the key bottleneck in RL training pipelines. Although Multi-To…
Real-Time Language Model Jamming: A Case Study for Live Music Accompaniment Generation
Bowen Zheng, Andrew H. Yang, Jiaqi Ruan +5
Language models (LMs) have become one of the most prominent paradigms in modern generative modeling. While making them faster has been the main focus of real-time deployment, speed…
SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training
Shengkun Tang, Zekun Wang, Bo Zheng +7
Structured pruning and knowledge distillation (KD) are typical techniques for compressing large language models, but it remains unclear how they should be applied at pretraining sc…
MaskTab: Scalable Masked Tabular Pretraining with Scaling Laws and Distillation for Industrial Classification
Bo Zheng, Yudong Chen, Zihua Xiong +4
Tabular data forms the backbone of high-stakes decision systems in finance, healthcare, and beyond. Yet industrial tabular datasets are inherently difficult: high-dimensional, ridd…
Accelerating Compound LLM Training Workloads with Maestro
Xiulong Yuan, Hongqing Chen, Jiaxuan Peng +16
Compound LLM training workloads-such as knowledge distillation and multimodal LLM (MLLM) training-are gaining prominence. These typically comprise heterogeneous components differin…