Hanrong Ye

Latest

• Attend Before Attention: Efficient and Scalable Video Understanding via Autoregressive Gazing
• Scaling Parallel Sequence Models to Vision Foundation Models
• OmniVinci: Enhancing Architecture and Data for Omni-Modal Understanding LLM
• Fast-SLM: Towards Latency-Optimal Hybrid Small Language Models
• GSPN-2: Efficient Parallel Sequence Modeling
• Scaling RL to Long Videos