Update README.md

README.md

@@ -2,6 +2,16 @@
 license: apache-2.0
 ---
 
-PyVision-RL: Forging Open Agentic Vision Models via RL
+## PyVision-RL: Forging Open Agentic Vision Models via RL
+
+This is PyVision-Video-7B-RL, post trained from Qwen2.5-VL-7B.
+
+```bibtex
+@article{pyvision2025,
+  title={PyVision-RL: Forging Open Agentic Vision Models},
+  author={Your Name},
+  journal={arXiv:2501.xxxxx},
+  year={2025}
+}
+```
 
-Reinforcement learning for agentic multimodal models often suffers from interaction collapse, where models learn to reduce tool usage and multi-turn reasoning, limiting the benefits of agentic behavior. We introduce \model-RL, a reinforcement learning framework for open-weight multimodal models that stabilizes training and sustains interaction. Our approach combines an oversampling–filtering–ranking rollout strategy with an accumulative tool reward to prevent collapse and encourage multi-turn tool use. Using a unified training pipeline, we develop \model-Image and \model-Video for image and video understanding. For video reasoning, \model-Video employs on-demand context construction, selectively sampling task-relevant frames during reasoning to significantly reduce visual token usage. Experiments show strong performance and improved efficiency, demonstrating that sustained interaction and on-demand visual processing are critical for scalable multimodal agents. Code, data and models are released at \url{https://github.com/agents-x-project/PyVision-RL}