TimeChat-Online: 80% Visual Tokens are Naturally Redundant in Streaming Videos
Open Access
- 27 October 2025
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 10807-10816
Abstract
No abstract availableAll Related Versions
Funding Information
- National Natural Science Foundation of China (92470205, 62176002)
This publication has 14 references indexed in Scilit:
- TopV: Compatible Token Pruning with Inference Time Optimization for Fast and Low-Memory Multimodal Vision Language Model2025
- VideoLLM Knows When to Speak: Enhancing Time-Sensitive Video Comprehension with Video-Text Duet Interaction FormatQeios Preprints, 2024
- An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models2024
- TimeChat: A Time-sensitive Multimodal Large Language Model for Long Video Understanding2024
- Chat-UniVi: Unified Visual Representation Empowers Large Language Models with Image and Video Understanding2024
- MovieChat: From Dense Token to Sparse Memory for Long Video Understanding2024
- VILA: On Pre-training for Visual Language Models2024
- Video-LLaMA: An Instruction-tuned Audio-Visual Language Model for Video Understanding2023
- Change DetectionAnnual Review of Psychology, 2002
- On the Surprising Behavior of Distance Metrics in High Dimensional Space2001