11 papers
DisagMoE: Computation-Communication overlapped MoE Training via Disaggregated AF-Pipe Parallelism
Zhichen Zeng, Chi-Chih Chang, Jiayi Wang +10
Mixture-of-experts (MoE) architectures enable trillion-parameter LLMs with sparsely activated experts. Expert parallelism (EP) is a widely adopted MoE training strategy, but it suf…
SRT: Accelerating Reinforcement Learning via Speculative Rollout with Tree-Structured Cache
Chi-Chih Chang, Siqi Zhu, Zhichen Zeng +5
We present Speculative Rollout with Tree-Structured Cache (SRT), a simple, model-free approach to accelerate on-policy reinforcement learning (RL) for language models without sacri…
Mesh-Attention: A New Communication-Efficient Distributed Attention with Improved Data Locality
Sirui Chen, Jingji Chen, Siqi Zhu +3
Distributed attention is essential for scaling large language models (LLMs) to long contexts, yet existing methods either have limited parallelism or incur high communication costs…
MegaScale-MoE: Large-Scale Communication-Efficient Training of Mixture-of-Experts Models in Production
Chao Jin, Ziheng Jiang, Zhihao Bai +16
We present MegaScale-MoE, a production system tailored for the efficient training of large-scale mixture-of-experts (MoE) models. MoE emerges as a promising architecture to scale l…
MegaScale-Infer: Serving Mixture-of-Experts at Scale with Disaggregated Expert Parallelism
Ruidong Zhu, Ziheng Jiang, Chao Jin +17
Mixture-of-Experts (MoE) showcases tremendous potential to scale large language models (LLMs) with enhanced performance and reduced computational complexity. However, its sparsely…
SwiftSpec: Ultra-Low Latency LLM Decoding by Scaling Asynchronous Speculative Decoding
Ziyi Zhang, Ziheng Jiang, Chengquan Jiang +5
Low-latency decoding for large language models (LLMs) is crucial for applications like chatbots and code assistants, yet generating long outputs remains slow in single-query settin…