Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning
Paper • 2607.18722 • Published • 35
The official organization of Tencent Hunyuan team
Scaling Native Multimodal Pre-Training From Scratch
Stale but Stable: Staleness-Adaptive Trust Regions for Stabilizing Asynchronous Reinforcement Learning