Penghui Qi

Revisiting Parameter Server in LLM Post-Training

SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning

Understanding R1-Zero-Like Training: A Critical Perspective

Optimizing Anytime Reasoning via Budget Relative Policy Optimization

PipeOffload: Improving Scalability of Pipeline Parallelism with Memory Optimization