June 22, 2026 ViCoStream: Streaming VideoLLMs Can Run Beyond 100 FPS with Stage-Wise Coordinated Inference 通过阶段式协调推理架构，实现流式 VideoLLM 在单 A100 GPU 上达到 134 FPS 吞吐量和低于 50ms TTFT video-understanding videollm streaming-inference token-dropping inference-optimization real-time-processing