by EvolvingLMMs-Lab · Agent Tool · ★ 43
ParaVT Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning Overview Long-video understanding is increasingly framed as agentic video reasoning: a large multimodal model (LMM) post-trained with reinforcement learning to invoke video-processing tools. Prior work in this line, including our earlier LongVT (CVPR 2026), dispatches tool calls sequentially — brittle to single mis-localizations, prone to multi-turn context drift, and linear in cost.
| Stars | 43 |
| Forks | 1 |
| Language | Python |
| Category | Agent Tool |
| License | Apache-2.0 |
| Quality Score | 70.7349419360204/100 |
| Last Updated | 2026-06-02 |
| Created | 2026-04-28 |
| Platforms | python |
| Est. Tokens | ~326k |
These tools work well together with ParaVT for enhanced workflows:
Explore other popular agent tool tools:
ParaVT is ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning. It is categorized as a Agent Tool with 43 GitHub stars.
ParaVT is primarily written in Python. It covers topics such as agentic-rl, grpo, long-video-understanding.
You can find installation instructions and usage details in the ParaVT GitHub repository at github.com/EvolvingLMMs-Lab/ParaVT. The project has 43 stars and 1 forks, indicating an active community.
ParaVT is released under the Apache-2.0 license, making it free to use and modify according to the license terms.