5 papers
Router Drift Control: Mitigating Off-Policy Degradation in Mixture-of-Experts Reinforcement Learning
Liangyao Jin, Zhixin Wang, Liming Liu +8
Reinforcement learning (RL) can become off-policy due to training-inference mismatch, mini-batch updates, and asynchronous training. The resulting mismatch is especially harmful in…
VenusRL: A Fully Disaggregated Agentic RL System with Priority Scheduling and Scalable Interaction
Mingjun Zhang, Yucheng Li, Menghao Zhang +15
Agentic Reinforcement Learning (RL) trains LLM agents through multi-turn interactions with external tool environments. Its multi-turn nature exposes two system-level bottlenecks un…
HetCCL: Enabling Collective Communication For Mixed-Vendor Heterogeneous Clusters
Yuejie Wang, Tao Chang, Yuanyuan Zhao +10
Training Large Language Models (LLMs) on heterogeneous clusters presents significant challenges for collective communication, as hardware from multiple vendors introduces diverse n…
An Efficient, Reliable and Observable Collective Communication Library in Large-scale GPU Training Clusters
Mingjun Zhang, Xiaohe Hu, Menghao Zhang +21
Large-scale LLM training requires collective communication libraries to exchange data among distributed GPUs. As a company dedicated to building and operating large-scale GPU train…
Expert-as-a-Service: Towards Efficient, Scalable, and Robust Large-scale MoE Serving
Ziming Liu, Boyu Tian, Guoteng Wang +15
Mixture-of-Experts (MoE) models challenge serving infrastructures with dynamic, sparse expert utilization, causing instability on conventional systems designed for dense architectu…