by ziqipang · Agent Tool · ★ 31
MR. Video: "MapReduce" is the Principle for Long Video Understanding Ziqi Pang, Yuxiong Wang [] [] Introduction What is the ultimate principle for long video understanding? Let's think about the most perfect criteria: digesting global contexts while perceiving local details simultaneously. For example, a question of "counting how many goals are there in the sports video" requires both global context (e.g., the overall scene) and local details (e.g., the specific action of scoring).
| Stars | 31 |
| Forks | 0 |
| Language | Python |
| Category | Agent Tool |
| License | MIT |
| Quality Score | 56.9527926618985/100 |
| Last Updated | 2026-06-18 |
| Created | 2025-04-10 |
| Platforms | python |
| Est. Tokens | ~14k |
Explore other popular agent tool tools:
MR-Video is MR. Video: MapReduce is the Principle for Long Video Understanding. It is categorized as a Agent Tool with 31 GitHub stars.
MR-Video is primarily written in Python. It covers topics such as agents, llm-agent, llms.
You can find installation instructions and usage details in the MR-Video GitHub repository at github.com/ziqipang/MR-Video. The project has 31 stars and 0 forks, indicating an active community.
MR-Video is released under the MIT license, making it free to use and modify according to the license terms.