Views
No views yet
| Checkpoint | Link | Note |
|---|---|---|
| Qwen2.5-VL-Base | link | Used as initial weights for training. |
| Robust-R1-SFT | link | Fine-tuned on Robust-R1 dataset |
| Robust-R1-RL | link | Fine-tuned with reinforcement learning on Robust-R1 dataset |
1@inproceedings{tang2025robustr1,
2 title={Robust-R1: Degradation-Aware Reasoning for Robust Visual Understanding},
3 author={Tang, Jiaqi and Chen, Jianmin and Wei, Wei and Xu, Xiaogang and Liu, Runtao and Wu, Xiangyu and Xie, Qipeng and Wu, Jiafei and Zhang, Lei and Chen, Qifeng},
4 booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
5 year={2026}
6}